AN ECONOMIC ANALYSIS OF SMALL WATERSHED PROJECT EVALUATION PROCEDURES THESIS FOR THE DEGREE OF PH. 1). mcmam sum; {NEW JOHN VONBRUSKA 197i LIBRAR" mulmnmmmmnmmm WWI L {NEWS 3 1293 00836 0210 Michigan State University I. This is to certify that the thesis entitled AN ECONOMIC ANALYSIS OF SMALL WATERSHED PROJECT EVALUATION PROCEDURES presented by John VOndruska has been accepted towards fulfillment of the requirements for Ph.D. degreein Agricultural Economics Mala.» JM Major professor Date February 1. 1971 0-7639 ! I i -__ VA 2 In evaluating pnjects, the Soil States Department 1 preeedures. The e1 are studied using t Presented and used . 53: estimating agri 553531 "it? to nude: 0 ”“353 and capital 1' In addition, ABSTRACT AN ECONOMIC ANALYSIS OF SMALL WATERSHED PROJECT EVALUATION PROCEDURES By John Vondruska In evaluating and justifying investments in small watershed projects, the Soil Conservation Service (SCS), an agency of the United States Department of Agriculture, employs numerous assumptions and procedures. The effects of changes in these procedures and assumptions are studied using the technique of sensitivity analysis. Two models are presented and used. One model is a systematization of SCS procedures for estimating agricultural (crop) benefits and is used to study benefit sensitivity to underlying crop—enterprise and hydrological variables. The second model is used to study the effects of change in SCS assumptions on benefit and cost timing, patterns and annual flow rates; interest rates; and capital investments for 12 Michigan projects. In addition, the historical background of the small watershed program is considered to help explain the dual emphasis on soil and water conservation, and water resource development, particularly flood control. FEderalvlocal costvsharing arrangements are detailed, including the inmortance of ACP (Agricultural Conservation Program) payments as an (Element of Federal cost. .Department of Agricultural and Congressional pnalicy preferenCes and their effect on project selection, planning and evaluation are also dz‘ cast ratio and alterna the conceptual standpc 'u‘hile there may b inith is most sensitiv data that are sensitiv 25:; SCS interest rat vi: the rate sets dif. evaluated the 12 projet Slizillion is obtainec anga single rate of 9 a: sun to a negative income at 52 inter Must. Other signif meat and cost timing “eject-credited f an“~ - lttt fann Income) ic tin" ' :drologlcal variah La :ezange; ( 1) the dire :v ,‘f‘v. v "v’ha ter damage reduq Litiina laggregate of . 9 its is 53 ObjectiVe th' :1 .D‘“! - “J'E‘ wage IQOUC John Vondruska evaluation are also discussed. Besides this, the SCS annual benefit cost ratio and alternative, noneSCS investment criteria are studied from the conceptual standpoint. While there may be some question as to which should he used and which is most sensitive, all of the studied investment criteria produced data that are sensitive to a host of agency assumptions and procedures. Using SCS interest rates (with as many as four rates per project and with the rate sets differing over the period 1959-68, when SCS originally evaluated the 12 projects), a lZ-project net present value sum of about $20 million is obtained, approximately equivalent to the sum based on using a single rate of 4%. Increasing the interest rate to 12% reduced this sum to a negative value, as would a 50% decrease in project—credited farm income at 5% interest, or a 100% increase in capital costs at 5% interest. Other significant changes include alterations in SCS assumed benefit and cost timing, patterns and cash flow rates. Project-credited farm income (the difference between with and without project farm income) is sensitive to changes in underlying crop-enterprise and hydrological variables. This hydrological sensitivity is significant because: (1) the directly affected category of agricultural benefits, floodwater damage reduction benefits, accounts for about onevhalf of the national aggregate of all project benefits; (2) hydrological data are much less objective than is sometimes thought; and (3) agricultural floodwater damage reduction benefits receive policy preference and emmhasis, even though they are but one form of projectncredited farm intmme, all forms of which depend essentially on increased crop output. AN EC WATERSHED ‘iit in Partia fl Departme. AN ECONOMIC ANALYSIS OF SMALL WATERSHED PROJECT EVALUATION PROCEDURES BY _ John Vondruska A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR or PHILOSOPHY . Department of Agricultural Economics 1971 he author would his helped make this . State Conservationist aw- itments for the stuc Pinning Party in Mich inlaid and discuss va Planing Party Leader; einsists; Russell Ba iii hydrologists. J oh: Slististions and comment E51333 Of this study. The author's gumé pra‘t‘i led many Construct {lat helped Shape the s Fuse helpful, that is an) Connor. ACKNOWLEDGMENTS The author would like to express his appreciation to several people who helped make this study possible. Robert S. Fellows, SCS Assistant State Conservationist in Michigan provided permission to use SCS in—file documents for the studied watershed projects. Members of the SCS Planning Party in Mfichigan offered helpful comments and took the time to explain and discuss various SCS procedures. They include Loren Oshel, Planning Party Leader; Arlo Benzmann, Justin Murray and John Okay, economists; Russell Baurle, Keith Bakeman and H. A. Amsterburg, engineers and hydrologists. John Okay was particularly helpful in providing suggestions and comments on various ideas and in reading various draft versions of this study. The author's guidance committee chairman, Dr. A. Allan Schmid, provided many constructive suggestions and asked pertinent questions that helped shape the study. Comments by other committee members were also helpful, that is, by Drs. Lester Manderscheid, Milton Steinmueller and Larry Connor. In addition, Dr. Ray Hoglund assisted in evaluating some crap enterprise data and in preparing such.data for use in chapters 5 and 6. Dr. Jim Beard provided similar assistance for lawn grass sod. LDrs. Ernest Kidder and George E. Merva reviewed chapter 5 from the hydrologist ' s viewpoint . In terms of manuscript preparation, Miss Jacqueline Kelly constructed true illustrative figures, Mrs. Mary Helen Ives helped improve the clarity ii of the text, and Miss 3- Hrs. Beverly Sager typt author, are on the star Farina Fisheries Servi. Particular apprec;1 rim provided encourage: intone of the text, and Miss Mary Vartabedian, Mrs. Ellen Rosenberg and Mrs. Beverly Sager typed the final copy. These peOple, as well as the author, are on the staff of the Division of Economic Research, National Marine Fisheries Service, U.S. Department of Commerce. Particular appreciation is expressed to the author's wife, Ellen, who provided encouragement and for the duration of this study, our income. iii LIST OF TABLES llSIOF FIGURES . . . tuner I. INTRODUCTION Agency-Comp Hypotheses 11. THE SMALL WAT Program Eme Public Law Conservatio lanning an Flood Versu. mumary III. THE SCS MODEL An Overview LIST OF LIST OF Chapter I. II. III. IV. TABLE OF CONTENTS TABLES . . . . . . . . . . . . . . . . . . FIGURES . . . . . . . . . . . . . . . . . INTRODUCTION . . . . . . . . . . . . . . . Agency-Computed Versus Base—Estimate Data Hypotheses THE SMALL WATERSHED PROGRAM . . . . . . Program Emergence Public Law 566 Conservation and PL 566 Planning and Coordination Flood Versus Drainage Problems Summary THE SCS MODEL . . . . . . . . . . . . . An Overview FWDRB and Hydrology FWDRB Estimation: Computational Steps Enhancement Benefits Estimation: Computa- tional Steps Project Costs Obtaining the Benefit Cost Ratio Data Inputs, Sources and Assumptions Net Project Effects Summary SOME CONCEPTUAL PROBLEMS . . . . . . . . Efficiency Criterion Rules Social Discount Rates Sensitivity Analysis n— A Prologue iv Page vi ix 12 47 90 Chapter ' V. PEDRB AXD HE Methodolog Antecedent Soil Moistl Annual and Statistical: Point-Area Monthly L05 Summary v1. CROP EHERPRI Methodolog.v All'CI‘OP’ P Analysis Alternative Alternative Multiple AS: Alternative Summary Til. INVESTMENT CR2 PATTERN AND Ll Methodology Ranking Pro j Chapter V. VI. VII. VIII. BIBLIOGRAPHX .. . . . . . . . . . APPENDIX FWDRB AND HYDROLOGICAL ASSUMPTIONS . . . . . ‘Methqdology Antecedent Soil Moisture Soil Muisture and Plant-Growth_Density Annual and Partial-Duration Series FWDRB Statistical Concepts and FWDRB Point-Area Rainfall Adjustments Monthly Loss Probabilities (PM's) Summary CROP ENTERPRISE ASSUMPTION . . . . . . . . Methodology All-Crop, Percentage Change Sensitivity Analysis Alternative Crop Prices Alternative Crop Yields Multiple Assumption Changes Alternative Cropping Patterns Summary INVESTMENT CRITERIA: BENEFIT AND COST TIMING, PATTERN AND LEVEL ASSUMPTIONS . . . . . . . Methodology and Investment and Criteria Ranking Projects Interest Rates Enhancement Benefits Analysis Cash Flow Timing and Instant Installation FWDRB Redefined Adverse Farm Income and Capital Cost Changes Summary INTEGRATION AND SUMMARY . The SCS Model Investment Criteria Sources of Possible Error Conclusions Recommendations Page 128 183 214 269 298 306 isle Payment of PI by Federal a Summary of Va Branch of Rh Cost Allocati Reaches 1 an Watershed . Outline of Be: Simplified Loss Value Cor Composite Acre Annual Probab] Obtaining Net Obtaining Lucn Shanty of Net Benefits ComPlJtation of Smary of Anf‘ Table 2.1 2.2 2.3 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 5.1 5.2 5.3 LIST OF TABLES Page Payment of PL 566 Small Watershed Project Costs by Federal and Local Project Investors . . . . 19 Summary of Various Kinds of Project Costs, North Branch of Mill Creek Watershed, Michigan . . . 21 Cost Allocation, Channel Improvements for Reaches l and 2, North Branch of Mill Creek Watershed . . . . . . . . . . . . . . . . . . . 27 Outline of Benefit Computational Categories, Simplified . . . . . . . . . . . . . . . . . . 49 Loss Value Computations, Corn for Grain . . . . 63 Composite Acre Value, Depth 1 . . . . . . . . . 64 Annual Probable Flood Damages . . . . . . . . . 67 Obtaining Net Returns (NR), Reach 1 . . . . . . 73 Obtaining LUCB Acreage, Economic Reach 1 . . . . 74 Summary of Net Returns and Enhancement Benefits 0 O C O O O O O O O C C O O O O O O O 75 Computation of EB from EB~100%, MILUB Data . . . 78 Summary of Annual Benefits and Costs . . . . . . 81 Alternative Runoff Curve Numbers (CN's) and Relative MR3 0 O O O I O O O O O O O O O O O 135 Growing Season Flood and Storm Probabilities ‘Compared, Michigan Data . . . . . . . . . . . . 144 TMonthly FWDRBAM, PM.and Related Data, Example PrOject O O O O O O O O O O O O O O O O O O O O 167 vi Table Monthly Runoi Index of FWDR Base Values f Approximate B Benefit Sensi Project, Mic] Benefit Sensil Project, Micl Effects of Alt Selected CrOp Effects of Alt Yields Effects of Alt Cropping Pat te FWDRB . k Invesmem Cri Table 5.4 5.5 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7.1 7.2 7.3 7.4 7.5 ~ 7.6 Page IMonthly Runoff-Potential Variations . . . . . . 170 Index of FWDRB for 12 PM—Set Assumptions . . . . 172 Base Values for the Example Projects . . . . . . 186 Approximate Benefit Response Coefficients . . . 187 Benefit Sensitivity Indexes, Mill Creek Project, Michigan . . . . . . . . . . . . . . . 190 Benefit Sensitivity Indexes, Tebo Erickson Project, Michigan . . . . . . . . . . . . . . . 191 Effects of Alternative Price and Cost Sets . . . 196 Selected Crop Price Data . . . . . . . . . . . . 198 Effects of Alternative Crop Prices, Costs and Yields 0 O O O O O O I I O O O O O O O O O O O 202 Effects of Alternative Prices, Costs, Yields, and Cropping Patterns . . . . . . . . . . . . . 208 Cropping Patterns and Crop Contributions to FWDRB O O O O O I I O O O O O O O O O O O O O 0 210 Investment Criteria Base Estimate Data, 12 Michigan PL 566 Projects, 1959-68 . . . . . . . 230 Twelve Michigan PL 566 Projects Ranked by Selected Investment Criteria . . . . . . . . . 232 Single Interest Rate Proxies and Internal Rates of Return, 12 Michigan PL 566 Projects . . . . 240 Net Present Value Sums, Selected Discount Rates, 12 Michigan PL 566 Projects . . . . . . 246 Changed EB Achievement Rates, 12 Michigan PL 566 Projects . . . . . . . . . . . . . . . . 252 Changed EB Achievement Rates and ACK Economic Lives, 12 Michigan PL 566 Projects . . . . . . 254 vii _——l _....\_ Table 7.7 Effect 12 Mic LS Undiscor Project L9 Comparis Capital Project Appendix Table L Project C 1959-68, 2' Capital I; Rates US( Factors 5 COmPUtati Benefits a Prolects, Names and Michigan L U'S- and M: Categoriz. Yield Data, .Chigan S Alternative Michigan Table 7.7 7.8 7.9 Effect of Altered Installation Timing, 12 Michigan PL 566 Projects . . . . . Undiscounted Net Cash Flows, Mill Creek PrOject O O O I O O O O O O O O O O O O O O 0 Comparison of Adverse Net Farm Income and Capital Cost Changes, 12 Michigan PL 566 Projects . . . . . . . . . . . . . . . Appendix Table 1. 2. Project Cost Data, 12 Michigan PL 566 Projects, 1959-68, Data in Dollars . . . . . . . . . Capital Investment Economic Life Data, Interest Rates Used by SCS for Capital Cost Amortization Factors and for Enhancement Benefits (EB) Computations, and EB Achievement Rate Data Benefits and Related Data, 12 Michigan PL 566 PIOJECCS, 1959-68 0 o o o o o o o o o o o o o 0 Names and Locations, 12 Studied and Other Michigan PL 566 Projects, 1957-70 . . . U.S. and Michigan PL 566 Project Benefits, Categorized, Data in Dollars . . . . . . . . Yield Data, Mill Creek Project and 1959-63 Michigan State Average . . . . . . . . . Alternative Crop Prices in Dollars, for MiChigan O O O O O O O O O O 0 viii Page 257 258 263 306 307 309 310 311 312 313 Figure 3.1 7.1 7.2 SCS Damage F r W Curves, M NPV Curves, M Alternative E; EB Growth Patt Figure 3.1 7.1 7.2 7.3 7.4 LIST OF FIGURES SCS Damage Frequency Curve . . . . . NPV Curves, Michigan PL 566 Projects NPV Curves, Michigan PL 566 Projects Alternative EB Growth Patterns . . . EB Growth Patterns for Four Michigan Projects Page 66 243 244 248 250 This is a study < snail watershed projefi farcers in overcoming not to deny the insta it to deny their prov pollution abatement a suggest that small va providing services th inlh'thigan and for t lite this study was related largely to se agency Plans across Although the sma afltiple Purpose in y nth agencies as the‘ elennessee Valley‘ individual Planning I! Latent 00 Sinare mi V at . a1 pIOJthS can 55 a CHAPTER I INTRODUCTION This is a study of government agency procedures for evaluating small watershed projects which are installed in rural areas to assist farmers in overcoming flood, drainage and irrigation problems. This is not to deny the installation of these projects in urban areas, nor is it to deny their provision of services for recreation, water supply, pollution abatement and other uses. In fact there is evidence to suggest that small watershed projects are becoming less oriented to providing services that primarily benefit farmers. This is true both in Michigan and for the United States as a whole. However, at the time this study was begun in 1967 agency plans for projects in Michigan related largely to serving agriculture and the same is true for agency plans across the country. Although the small watershed program has the potential of becoming multiple purpose in character, like the large watershed programs of such agencies as the Corps of Engineers, the Bureau of Reclamation and the Tennessee Valley Authority, the legislated constraint limiting individual planning units (projects) to watersheds of 250,000 acres (about 400 square miles) may offset some of this potential. Of course, several projects can be for contiguous watersheds, so long as each has a separate watershed work plan. The Soil Conserv. responsibility for pl. so far as the Federal oflgriculture (ESQ-i) services marshaled to rent provides project the application of co: the sense of Federal a financial assistance I and individual land-0v Eeing more favorable y lending institutions) installation of small control counterparts , watershed Projects f0 Opera Ltd and maintain 2 The Soil Conservation Service (SCS) has been delegated primary responsibility for planning and developing small watershed projects so far as the Federal Government is concerned. Yet, other Department of Agriculture (USDA) agencies help provide the package of Federal services marshaled to assist local beneficiaries. The Federal Govern- ment provides project planning assistance, technical assistance in the application of conservation practices, investment underwriting (in the sense of Federal assumption of part of the investment cost), and financial assistance (in the way of loans to local project sponsors and individual land-owning beneficiaries, with the terms of the loans being more favorable than for loans available from regular, commercial lending institutions). These Federal services relate to the initial installation of small watershed projects. Unlike large basin flood control counterparts, as planned by the Corps of Engineers, small watershed projects for flood prevention (and other purposes) are operated and maintained by local project sponsors rather than by the Federal Government. The Soil Conservation Service (SCS), as its name implies, views its mission in the conservation framework. SCS sees the small watershed program in the same light. Both the agency and the program began in 1933 under authority of the National Industrial Recovery Act. This emphasizes the economy stimulating aspect of the program, common to other public works programs. Under this 1933 legislation the program would have been temporary in nature. It was given permanence under the Flood Control Act of 1936 as an upstream, small watershed counter- Part of the Corps of Engineers' downstream, large basin program. Except for work in eleven river basins for which general plans were approved in the 1944 F Flood Control Act was Protection and Flood P 336 Congress, 68 Statu authority imposed size could be planned and C if it is intended to I prevent overflow onto congressional control to the agricultural dc- projects with larger 1 23,000 acre feet limi“! Incidentally, the mov. agricultural funding . Million in the USE Soall watershed proje to test to ' the “Inmate be an operatiOnal sin 1.. 8CS U33e~bo d . I ‘It ’Seml‘put Work heets ( 3 approved in the 1944 Flood Control Act, SCS authority under the 1936 Flood Control Act was repealed by the passage of the Watershed Protection and Flood Prevention Act of 1954 (Public Law 566, 83d Congress, 68 Statutes 666), commonly called PL 566. This new authority imposed size limits on watersheds for which single projects could be planned and on reservoir size (flood detention capacity if it is intended to temporarily hold back floodwater so as to prevent overflow onto the floodplain). PL 566 also transferred congressional control and funding away from the public works domain to the agricultural domain, except in the matter of approval of Projects with larger reservoirs (over 4,000 acre feet and up to the 25,000 acre feet limit for single reservoirs in PL 566 projects). Incidentally, the move to agricultural committee control and agricultural funding actually began in 1953 with the approval of $5 IIlillion in the USDA budget specifically for a system of pilot Small watershed projects located across the country, reportedly to test the "ultimate worth" of the small watershed approach. Such a goal may be questioned, because the small watershed program had been operational since 1933. The Michigan small watershed projects considered in this study were planned by SCS under authority of PL 566, as amended. In all, 12 Projects are studied in chapter 7, although one project serves as the primary example in other chapters. The source of data consists of the SCS watershed work plans (short, mimeographed, loose~bound, semi-public documents) and the SCS in-file Planning Party work sheets (of detailed computations, called in-file SCS d oQ-llments or documentation to distinguish them from the work plans). lithough the first Mic was planned in 1957, E to the SCS central st. examination by the and Contrary to the agric' projects in Michigan, flood control for the of Michigan's lower p be Mackinaw Bridge c ye: able to make avai “it: . met propect group i tans primarily recre; process of being rev‘ Th I e 12 studied l 'D. ' . ms of their agrict of - ' cgrlcultural bene' Although the first Michigan small watershed project (under PL 566) was planned in 1957, SCS in-file documentation has been forwarded to the SCS central storage unit and it was not available for examination by the author in the SCS Planning Party offices. Contrary to the agricultural orientation of most other PL 566 projects in Michigan, this first project was geared to providing flood control for the small city of Cheboygan, located at the tip Of Michigan's lower peninsula, near the Straits of Mackinaw and the Mackinaw Bridge crossing to the upper peninsula. SCS was not yet able to make available the in-file documentation for the Maple River project group in central Michigan, because the plans for this primarily recreational group of projects were still in the Process of being reviewed and revised within the agency. The 12 studied Michigan projects were justified largely on the basis of their agricultural benefits. Nationally, the composition of agricultural benefits differs in emphasis, comparing the relative importance of agency—defined categories of benefits for Michigan and the United States as a whole. Nationally, _FWD_RB (floodwater damage reduction benefits) are more important than E\B~ (enhancement benefits). Both kinds of agricultural benefits relate to the project-credited increase in net farm income. The separation and emphasis of FWDRB relates to the policy directives issued by the House Agricultural Committee and by the D epal-Irtment of Agriculture, as discussed in chapter 2. SCS procedures for computing PL 566 project agricultural benefits are described in chapter 3 using numerical data for an e“ample project. The procedures are formulated into a mathematical aodel that has been wn used in chapters 5 am are not presented, tho chapter 3 provides a l ahost of hydrologica general economic vari cation of SCS computa crop prices, costs, y antage armual benef i development of capita scrks of improvement I P .urtnermore, the empE' secondary agricultural agricultural areas) a 0i this study. The author's SC [fies certain data J a p . cononnsts develop crl Production item . J of cost adjuStment f; 5 model that has been written into an operational computer program, as used in chapters 5 and 6. While the details of this computer program are not presented, the mathematical formulation of the SCS model in chapter 3 provides a basis for posing questions about the effect of a host of hydrological, watershed, crop enterprise and other more general economic variables. The author's SCS model allows the dupli- cation of SCS computational steps leading from such data as individual crop prices, costs, yields and planting patterns to the production of average annual benefits and the SCS annual benefit cost ratio. The development of capital cost and operations cost data for the project Works of improvement is not considered in the author's SCS model. Furthermore, the emphasis is entirely on agricultural primary benefits. Secondary agricultural benefits, redevelopment benefits (even for agricultural areas) and other kinds of benefits are outside the scope of this study. The author's SCS model of chapter 3, as used in chapters 5 and 6, takes certain data aggregates as given. For example, even though SCS e'Qotlomists develop costs for various crop enterprises from individual production item costs, the SCS economists' crop enterprise cost a8St‘egates are taken as data inputs for the author's SCS model. Crop costs can be adjusted in the SCS model of chapter 3 by the application of Cost adjustment factors. Altering factor of production combinations w ould be more difficult, requiring computation of entirely different Crop cost aggregates. This sort of change is more of a computational bl“rden than may appear at first glance. For the development of EB (enhancement benefits , estimating FhDRB (“0 data on a monthly bas. In addition to t': corputational procedu formulations are used and cost data for all and this is reflected short-cut procedures with access to a big}.- is reduced. Specifyi period is a conceptua variations from SCS SCS-assumed benefit ' in chapter 7. Also, studied. In additio: mesrnent criteria ' (enhancement benefits) annual crOp production costs are used, but estimating FWDRB (floodwater damage reduction benefits) requires cost data on a monthly basis. In addition to the author's model of SCS agricultural benefit computational procedures of chapter 3, a second set of mathematical formulations are used. SCS economists do not actually compute benefit and cost data for all years in the evaluation period (t = l, ..., T), and this is reflected in the SCS model of chapter 3. They use some short-cut procedures adapted for use with desk calculators. However, With access to a high speed digital computer, the burden of calculation is reduced. Specifying net cash flows for all years in the evaluation Period is a conceptually preferable approach, given the type of Variations from SCS assumptions and procedures studied in chapter 7. SCS~assumed benefit and cost timing, pattern and rate data are changed in Chapter 7. Also, the effect of different discount rates is Studied. In addition to the SCS annual benefit cost ratio, other in‘VEStment criteria are used, notably the net present value and internal rate of return. Data computed for altered assumptions for these inVestment criteria are then compared to analogous data for base- estilllate assumptions. Chapters 5 and 6 are complete studies in themselves, but they may be Viewed as providing a rationale for changing the SCS—computed ‘benefit data in chapter 7. Chapter 5 considers the relationship bet"Ween FWDRB (floodwater damage reduction benefits) and underlying hydrological variables and assumptions. Chapter 6 considers the r elationship between all categories of SCS-computed agricultural b enefim and underlying crop enterprise variables (such as yields, prices, costs and P1‘ Sand 6 determine th‘ one of the data inpU' chapter 7. Chapter 4 is a . investment rules and project evaluation. literature for the f: provide a rationale 1 discount rate data as Chapter 8 is an Agency-CC Where the compal peso-estimate data a; ‘I I other hand, the auto essentiall uplicat Chapter 3 mid prices, costs and planting patterns). The variables studied in chapters 5 and 6 determine the level of project-credited farm income, which is one of the data inputs for the investment criteria formulations of chapter 7. Chapter 4 is a discussion of national efficiency-criterion investment rules and some conceptual problems related to their use in project evaluation. This chapter is not intended to be a review of literature for the field of public investment analysis. It does provide a rationale for studying both benefit and cost data and discount rate data assumed in agency evaluations. Chapter 8 is an integration and summary of previous chapters. Agency—Computed Versus Base-Estimate Data Where the comparison is relevant and plausible, the author's baSe—estimate data approximate analogous SCS data. Generally, base- estimate data reported in chapters 5-8 are benefits, benefit cost ratios or numerical expressions for other investment criteria (such as Pr()l'iect net present values, internal rates of return or present value benefit cost ratios). Except for undiscounted net cash flow data, this final-step data is the only kind developed for chapter 7. 0n the other hand, the author's detailed SCS model, used in chapters 5 and 6, essentially duplicates the step-by-step computations explained in chaPter 3. For purposes of analysis with this detailed, sequential model of SCS procedures, and to assist in the initial debugging of the Q0111Pinter program for it, intermediate data was added to the computer IMtillt-out sheets. Very little of this detailed, intermediate data is DreSented in chapters 5—8. Nevertheless’ h neonate: out?“ data nade t0 Precise1y dun eventually to the bag as can be seen by 90‘ Sorth Branch of Mill and by comparing data projects, the author' differ slightly. Age- the Mill Creek projec total of $282,065. 8 input data, they are benefit cost ratios 0 computations of chapt ratio for the Mill Cr Why do the resull iztemediate data for the author found it n C! -5 °-° ' Had the scs e 8 Nevertheless, having the SCS step—by-step data and analogous computer output data proved quite valuable. Considerable effort was made to precisely duplicate the SCS results at every step leading eventually to the base estimates reported in chapters 5 and 6. However, as can be seen by comparing data in Tables 3.9, 6.1, and 7.1 for the North Branch of Mill Creek project (or Mill Creek project, for short), and by comparing data in Appendix, Table 3, for the 12 studied PL 566 projects, the author's base-estimate and agency-computed results differ slightly. Agency-computed total annual average benefits for the Mill Creek project are $286,020, compared to the base-estimate tOtal of $282,065. Since average annual costs are used as SCS model input data, they are the same in both cases, $40,520, leading to benefit cost ratios of 7.06/1 and 6.96/1, reSpectively. The simplified computations of chapter 7 provide a second base-estimate benefit cost ratio for the Mill Creek project, 7.35/1. Why do the results differ? First, because the SCS economist used intermediate data for some crops in the Mill Creek project evaluation, the author found it necessary to develop SCS model input data for these Crops. Had the SCS economist gone back to the same stage in the chain of computations, he would have obtained different intermediate data than he actually used. What the SCS economist did resulted in a coIlsiderable saving of time and did not affect the results Significantly. Secondly, it appears that the SCS economist made some “Li not“ computational errors. Because the SCS economist's and the an th013's step-by-step computations agree precisely for several creps, and because the diffe sill be assumed that an accurate represent To reiterate, ft economist's total an: ratio, 7.06/1 are qui and 6.96/1, reSpecti‘; relatively larger, bu For the base-est required for the ques base-estimate data d; cost ratios of the so; the author also uses t0 the data for non-Sp 333d 7, SCS economilI for all years in the l=100, depending 0; all years for Other aPie‘l'oach a o 130 In for, re:ios. ~ . I IS consis and because the differences can be reconciled for other crops, it will be assumed that the author's SCS model of chapter 3 is in fact an accurate representation of SCS procedures. To reiterate, for the North Branch of Mill Creek project, the SCS economist's total annual average benefits, $286,020 and benefit cost ratio, 7.06/1 are quite close to author's base-estimate data, $282,065 and 6.96/1, respectively. Differences in benefit sub-categories are relatively larger, but they have been reconciled, as indicated here. For the base-estimate data of chapter 7, another answer is required for the question: Why do the agency-computed and the author's bEisen-estimate data differ? Comparison is possible only for benefit cost; ratios of the SCS type (shown in Appendix, Table 3), although the author also uses the term base estimate in chapter 7 when referring to the data for non-SCS investment criteria. As indicated in chapters 3 and 7, SCS economists do not actually compute various cash flows for all years in the evaluation period (t = l, ..., T, where T = 50 or T ‘3 100, depending on the project). Since cash flows are needed for all Years for other investment criteria, the author decided to use this apPreach also in formulating approximations of the SCS benefit cost ratios. This consistency of approach has certain conceptual advélntages that will not be discussed here. However, this departure from strict use of SCS procedures should not give rise to any Significant differences in results. Examination of the two sets of benefit cost ratios (in Appendix, Table 3) will show that differences are generally slight, usually less than :5%, some of which could be e xplained by rounding effects in the chain of computations. In other cases, Mutational and P“ evaluated over the pe to discuss these 8W performed the origins had been transferred positions within the in most agencies of In a few cases, changed midway throu Given the long chain prfieet investment, possible of what has usually quite system agency for purposes “it SYstenatic of ; for the author to d" tional arrangements plan for the Projec To . surmiarlze, 10 In other cases, it appeared that the SCS economists had made some computational and procedural errors. Since the 12 projects were evaluated over the period 1959—68, it was not possible for the author to discuss these apparent errors with the particular economists who had performed the original evaluations. That is, some of the SCS economists had been transferred from the SCS Planning Party in Michigan to other positions within the agency, transfers being a common personnel practice in most agencies of SCS size. In a few cases, it appeared that project plans may have been changed midway through the evaluation. This entails recomputation. Given the long chain of computations involved in evaluating a PL 566 Project investment, one can appreciate the desire to salvage as much as Possible of what has already been done. While SCS economists are usLlally quite systematic in their calculations—-a requirement of the agency for purposes of review--a change in project plans can upset the “1081: systematic of people. In such cases it was sometimes difficult for the author to determine which SCS data, assumptions and computa- t1011211 arrangements had actually led to the results shown in the work plan for the project. To summarize, the author's base estimate data correspond quite clos‘ely to analogous. data computed by the SCS Planning Party economist who performed the original project evaluation. The rather insignificant differences may be due to errors of computation by either the author or the SCS economist. The precise agreement for parts of the Mill Creek ptoject computations and the closeness of the final results lends Qtedence to the assumption that the SCS model of chapter 3 is in fact an accurate representation of SCS evaluation procedures. This assumption is also second project, the author's and the SC Table 3, for all 12 that the two compute Furthermore, it is p benefit cost ratios cnapter 7 are accura Two hypotheses Hypothesis 1 is other investment cri to nmnerous underlyi and procedures relar variables having t o for project costs a Hypothesis 2 ml Specify values for the apparent worth I ptlons . 1 333‘ ll assumption is also affirmed by application of this SCS model to a second project, the Tebo Erickson in chapter 6. The closeness of the author's and the SCS economist's annual benefit cost ratio (in Appendix, Table 3, for all 12 studied projects) is taken as prima-facie evidence that the two computational approaches are equivalent in results. Furthermore, it is presumed that the agreement of these two sets of benefit cost ratios means that the SCS data and assumptions used in chapter 7 are accurately represented. Hypotheses Two hypotheses have been used as guides in this study. Hypothesis 1 is as follows: SCS annual benefit cost ratios and Other investment criteria data for PL 566 projects are quite sensitive to numerous underlying assumptions and procedures. These assumptions and procedures relate to crop enterprise, hydrological and other Variables having to do with the timing, pattern and achievement rates for project costs and benefits. Hypothesis 2 may be stated as follows: If one is willing to SPecify values for some of these variables, it is possible to alter the apparent worth of PL 566 projects considerably from that based on SCS assumptions. It is not the purpose of this study to specify what assumptions should have been used or what assumptions should be used in SCS evaluations. However, it is intended to show how possible alternative assumptions would affect the apparent worth of PL 566 1)I.°:Iects. Indeed, the matter of assumptions is one over which reason- able men disagree. This does not detract from the point that what a‘lennptions are used can affect investment rates and the welfare of People . This chapter is of the PL 566 progran Public Law 566, conse flood versus drainage The SCS small we mivater resource It Primarily at overcom, problems . CHAPTER II THE SMALL WATERSHED PROGRAM This chapter is intended to briefly survey the nature and context of the PL 566 program, and the topics include: program emergence, Public Law 566, conservation and PL 566, planning and coordination, flood versus drainage problems, and summary.1 Program Emergence The SCS small watershed program has dual roots in the conservation and water resource legislation of the 1930's, legislation then aimed Primarily at overcoming severe and widespread income and unemployment problems. \ References for this chapter include: Robert J. Morgan, Governing §5¥££_£Bzgservation (Baltimore: Johns Hopkins Press for Resources for :he Future, 1965); R. Bernell Held and Marion Clawson, Soil Conservation W (Baltimore: Johns Hopkins Press for Resources for the uture, 1965); Charles. M. Hardin, Food and Fiber in the Nation's Politics \ 0n F , Vol. III of Technical Papers, National Advisory Commission (,, 00 USGPO, 1967), esp. sec. 4 ThSOil Conservation Programs"); Luna B. Leopold and Thomas Maddock, WControl Controversy (New York: The Ronald Press, 1954); 1967’ SCS, The Watershed FrotectionHandEook (Washington,Da.C.: SCS, Prot, hereafter, fl; USDA, SCS, Economics Guidefiforj Watershed W51 and Flood Prevention (Washington, D.C.: SCS, 1964), u Zeafter, Economics Guide. The last two references are periodically hp ated, by page or section, and this will be indicated; also, both :ve now been issued in several editions, and earlier editions will clted if relevant. d and Fiber (Washington, D.C.: 12 Conservation and SCS Partly because to dramatize the pro became the head of o shortly after the in It is widely be 1935 provided t servation Servi as an emergency unemployment ur. was both SecretI the federal wor' | 139- Program was pop; 450 ccc (Civilian cC 13 Conservationfiand SCS Partly because of his efforts in the late 1920's and early 1930's to dramatize the problems of soil erosion, Hugh Hammond Bennett became the head of one of the many new agencies organized in 1933, shortly after the inauguration of President Franklin D. Roosevelt. It is widely believed that the dust storms of 1934 and 1935 provided the impetus for initiating the Soil Con- servation Service; in fact, erosion control was started as an emergency federal public works project to relieve unemployment under the direction of Harold L. Ickes who was both Secretary of the Interior and administrator of the federal works program.2 The program was popular with congressmen and by 1936 SCS was directing 450 CCC (Civilian Conservation Corps) camps and 151 conservation demonstration projects.3 CCC and WPA (Works Progress Administration) peeple did most of the work, although some labor and material was provided by farmers. Five-year farmer contracts were used, and they Serve as a precedent for present day conservation—practice agreements (not contracts) in the SCS-PL 566 context. Congress passed PL 46 in 1935, transferring the Soil Erosion Service (SES) from the Department of Interior to the Department of Agri- Culture, and renaming it the Soil Conservation Service (SCS). PL 46 is the basis for present day SCS conservation planning and technical \— 1933 2Quoting Morgan, pp. 1-2. The National Industrial Recovery Act of See authorized soil erosion control work to help relieve unemployment, pr Hardin, p. 146. SCS and the conservation—demonstration projects Ogram were given permanence under PL 46 of 1935 (see note 4), although ase projects were no longer funded by Congress after World War II. 3 Morgan, p . 42 . assistance especially are funded largely by acceleration funding projects is also prox in the form of ACP (E administered by the (ASCS, not SCS). as programs. Its begin Allotment Act of 193 atendment to PI. 46, Agricultural Adjustl‘. control over agricul involved both in con within USDA, and so: be noted.5 W To use an SCSe Program essentially ml the downstream ) 14 assistance especially to farmers.4 In PL 566 projects these services are funded largely by PL 46 appropriations, with supplemental acceleration funding under PL 566. Similar assistance for PL 566 projects is also provided by the Forest Service. Financial assistance in the form of ACP (Agricultural Conservation Program) payments is administered by the Agricultural Stabilization and Conservation Service (ASCS , not SCS). ASCS also administers farm price and income support programs. Its beginning traces to the Soil Conservation and Domestic ' Allotment Act of 1936, an act passed by Congress technically as an amendment to PL 46, after the Supreme Court declared the original Agricultural Adjustment Act (AAA) unconstitutional (for reasons of control over agricultural production). Thus, there are several agencies involved both in conservation and in small watershed projects, all Within USDA, and some amount of historical conflict between them should be noted. 5 good Control Protects and SCS To use an SCS explanation, the small watershed or upstream PIOgram essentially fills the gap between the on-farm (ASCS and SCS) and the downstream, large-basin programs. SCS did some small watershed work even before the passage of the 1936 Flood Control Act, probably in the form of conservation demonstration projects, but most SCS work outiside of PL 566 traces to this 1936 Act. \ 4 ‘U S Sec. 1, Public Law 46, 74th Cong., 1 Sess., 49 Stat. 163, 164 (16 E; .C. 590ae590f); hereafter, PL 46. This act is called both the Soil Osion and Soil Conservation Act of 1935. 5 Morgan, pp. 144-157; Hardin, pp. 154—164. 15 The 1936 Flood Control Act is often cited nowadays as benchmark legislation, but it should be noted that it was then intended in part to overcome Depression unemployment and income problems. It approved work broadly planned in the Corps of Engineers' "308" reports, made under the auspices of the 1927 and 1928 Rivers and Harbors Acts. These reports were by no means project plans or proposals, but they were the most complete and comprehensive river basin studies then available. These reports also served as a basis for the Tennessee Valley Authority Act of 1933. The Flood Control Act of 1936 states: 1. That flood control is a proper federal function and that the federal government should improve or participate in the improvement if the benefits to whomsoever they may accrue are in excess of estimated costs. 2. That a flood control program is justified if the lives and social security of people are otherwise adversely affected. Needless to say, these provisions have prompted considerable debate. The same can be said for the 1938 Flood Control Act that went another SteP and removed local responsibility for financial participation in corPsrbuilt reservoirs. Under the 1936 Act, local governments had to prov1de land, operate the completed project, and free the Federal government from responsibility for damage suits in connection with projects. These requirements now apply only to so-called "local" works (meaning levees and channel improvements) in the Corps, flood control context, bet siiilar to those for. Until repeal by" Flood Control Act. SCS survey reports l l944 Flood Control : survey upstream are. dmnstream surveys. 'rlorld War II. Small watershet Control Act, 1953 PI authority. The W By 1953 the SCI ‘iEar. SCS had prep to Congress that ye 16 control context, because of the 1938 Flood Control Act;6 they are similar to those for SCS work under PL 566. Until repeal by PL 566, SCS had survey authority under the 1936 Flood Control Act. Work is still in progress on projects outlined in SCS survey reports for eleven minor river basins, as approved in the 1944 Flood Control Act. SCS had been given authority in 1937 to survey upstream areas of basins then authorized for Corps of Engineers downstream surveys. Little SCS construction was funded until after World War II. Small watershed projects may be built under PL 566, 1944 Flood Control Act, 1953 Pilot Watershed appropriation or possibly other authority. fie 1953 Pilot Watershed Appropriation By 1953 the small watershed program had been underway for several year. SCS had prepared about 50 surveys, some of which were submitted to Congress that year; and plans existed for basins ranging in size \ 6Leopold and Maddock, pp. 83-104. 81 In 1927 Congress directed the Corps and the Federal Power Commis- rich to inventory the hydroelectric power potential of the nation's tiVet‘s. The list of streams was the result of Congressional authoriza- On Of 1925 for the two agencies, and it was published in House Cocmnent 20_8_ of the 69th Congress; hence, the name "308" reports. in tlnection with the surveys of these rivers authorized, in.1927. See (:2 MCreell, Our Nation's Water Resourcesj-Policies and Politics 13880, Illinois; The Law School of the University of Chicago, 1956), pp, 42 and 67. Also, see Roland McKean, Efficiency in Government 1h£‘L‘E8h‘Systems Analysis (New York: John Wiley and Sons Inc., 1958) p . 18. Y _ , i - . ’ 9 from 14 to 53,000 E niles for PL 566 pr upstream-downstrear $5 million was appr part of the 353.931 andlladdock cite a a"pilot plant" of value of the upstre oeen mdervay since Despite contrt shortly before elect not required unless feet, the maximum be Except for Si; 0: the multi‘PUI'Pose and the Corps of En Federal finan 17 from 14 to 53,000 square miles, compared to a maximum of 400 square miles for PL 566 projects. However, because of the big-small dam, upstream-downstream controversy, and possibly for other reasons, $5 million was appropriated in 1953 for SCS pilot watershed work, as part of the agricultural rather than flood control budget. Leopold and Maddock cite a House report indicating that Congress had in mind a "pilot plant" of 50 demonstration projects to test the ultimate value of the upstream work, although SCS small watershed work had been underway since even before the Flood Control Act of 1936.7 Public Law 5668 Deepite controversy, PL 566 passed Congress in the summer of 1954, shortly before elections. Public works committee approval is not required unless a project has a reservoir exceeding 4,000 acre feet, the maximum being 25,000 acre feet, as will be discussed shortly. Except for size restrictions, PL 566 has emerged as an analog of the multi-purpose authorities granted to the Bureau of Reclamation and the Corps of Engineers. Reservoirs can be constructed for a variety of Purposes. Federal financing of PL 566 flood-prevention construction is not ctuite the equivalent of Corps of Engineers' cost—sharing rates for flood- cOfflirol reservoirs which are financed entirely by Federal funds. \ 7 189 See Leopold and Maddock, pp. 208—210 and 230-232; Morgan, pp. 179.. 8 L The Watershed Protection and Flood Prevention Act of 1954, Public aw 566, 83d Cong., 2d Sess., 68 Stat. 666. Thus, PL 566 reserr flood protection not local interests mus Under PL 566, Feder control purposes, a Federal, local or 3 Classification eventual Federal-lo: determining loans ar for site preservatio ”a“! supply. Site Prior to the initiat. v-t ‘ er sopply "advance to the first usage 0 ”an to 302 of the cos Without any other fo Spcnsors' share of p and they are made by it the so~called Fed 18 Thus, PL 566 reservoirs are on a par with Corps of Engineers local flood protection works (levees and channel improvements), for which local interests must provide land and eventually Operate the project. Under PL 566, Federal cost-sharing rates are lower for non—flood— control purposes, and some functionally categorized costs may he Federal, local or shared, as shown in Table 2.1. Classification of project costs is important in determining the eventual Federal-local cost-sharing ratios and to some extent in determining loans and advances that may be made. Advances may be made for site preservation and the provision of industrial and municipal water supply. Site preservation advances must be repaid with interest Prior to the initiation of construction. Municipal and industrial water supply "advances" are interest free for up to 10 years (i.e., prior to the first usage of the new supply), and they may be used to finance HP to 30% of the cost assigned to this purpose (to be born locally, Without any other form of Federal assistance). Loans to cover the local Sponsor-3' share of project costs may range up to five million dollars, and they are made by the Farmers Home Administration for up to 50 years at the so-called Federal long-term bond interest rate.9 \ 9mm, sec. 103.041; PL 566, sec. 8. at 4 gegfrding interest rates, project costs (SCK only) are amortized at 13.11754 in fiscal 1970, and enhancement benefits. are computed local 8. rate (see chapter 7). However, for fiscal 1970 FHA loans to loan sponsors are made at 3.342%, and soil and water conservation ~ S'r-to. eligible farmers are to he made at 5% for 40. years; see USDA, 1976)LHAg_eampniet ?A‘7°5» revised August 1969 (Washhngton, D.C.: usopo, kHHWDMD WUCQUWMMMM!WWWMNMWWMIIIIFlII .U—tcu-k. QC—lrhdunu Luna‘s—Chm mucm Aura :nv mnumwv Ode ..Hnm EON...“ wufimwnw \fl.~.mhmwvw umOU wzu MO Now usonfl H0>OU RWE AU< kHuhmm AU< QH> COHUWUHHQQV .EOU< mam 20¢ on ouuoo ucmEumwhu tcmw UQUGSOU wfl %HCO OHUQH umOUIUHWCCUD hOUGHUESS hmucm ho QUHUUQHQ COHUW>hNWCOU Uflfim UHQQ. fimwfl 0300 NW. UCQEEWU kfififlufid kfidfihflhflh :OfiuflUHMflhmfiHU :OWuUCJH “00 l.. aupflgfllnunw U Chafinvuunh awn-23h 30 .63 H. ~ 333.... 0CD «Han HA...- 4 Coo—:33; . N. - «..1 new 521.? .flLOUEQ>CH dUflfiOhk dflUiJ USE dihfiflfih in [4130 .Amoma zoom: .oomfi>ou .ZOUm ca moomumammm momv Nco.lmeo.maa .moom was .Amumoo uomuucou mo .aea an cosuaaHUfiuums_mom .sema .amu .emma>muv Hmo.HOH .omm .Ammumuv Hmo.moa .omm .mm: .oom gm "muuaom mcmoa no uxou mom mmumou ommsuo>o com um :uoohouelaoa: mum mumoo waacamam poumnm woumnm mamoa .wafiacmae ”mumoo Honuo .oofi>om mom was Auoummoumzu moooam was woumm one whom» m umuHmV mdoauooemafi mum pom ueooxo .HmUOH Nooa haemoz oaom Azoumv oocwamucfimal.eo .HHm oofi>oue cu mow mmmmmmw >ma whomsoem Hmooa moma moafim :.Hmowuoao paw Hmwoa .m>HumuumHaHabm: .muomaoem Hmooa “moofi>uom Hmuaaaoou vowfi>one mom moma Hausa woumcm woumnm muomuucoo mo .afiao< .AHmUOH Nooav haeesm nouma Hmfiuumaoca was Hmeaowcsa cam A%Hao maeaflm mom How .oom Non ou adv omHHoHHsranm mam coaummuoou mom unooxo .Hmooa NOOH maumoz. oaom mama .Aamooa Nooav haeesm umumz Hmwuumsoaa was Hoe muuaoaasa com A.umm Mom ou adv moauwafiomm oMHHoaazlanm new aowumouomu ofimmn anafiofia pom ueooxo ..vom Nooa oaom saumoz moUH>uom mogumaamumcH .Hmooa NOOH ma nears magnum youm3 Hmfiuumsocfi new Hmefiofiana unouxm .momoeusm monuo mom .oom Non cu m: mooeuoo>oue vooam pom .vmm NOOH nonmam woumsm AneuumuuGOUV coauoauumaoo .zoom mumou Amxuoa waauooafiwco amouum pom Mom mm oeuwu umoonufimocoo mo ecumaaeoaov nouam Icfimav ouammoa Honouoauum .HmUOH NOOH HH< oooz monocouaamalaOfiumuomo .oom 4m .oumuoaooom ou mossy “meson ooa>uom umouom use Amaco momv we gm Boom nwmm maoumm kaamamb moomumwmmm Hmofiosooa .umoo man no Non usonm uo>oo ems mo< zauumm mo< mw> :OHuMUfiHee< .zoo< use Mo< mm mumoo uaoaumouu mama vmuaaoo ma mama oeumu umooluamocon noumuoasa Houam no oofiuomue coaum>uomaoo ofimm panel. aoufi umoo mo narrlnrrrlrolllrrrlwmmsaoo mfifimooq xfifimumemm cofiumUfiMflmmmHu eofiuucam mum I'lllllllullllllllllllllllll 1|||||||II|||I5||||||II\|I\I|'I|I|II\‘I‘|II||“ axe .H.N magma .muoumm>aH nomfioum Hmoou can snowman so mumoo noohoum eonmumnma Hanan com am we new m 20 Federal—Local Cost-Sharing Ratios: A Case Example A case example may he the best‘way to illustrate alternative cost— sharing ratios for PL 566 projects, although the resulting ratios may not be representative of the program in Michigan or nationally. National data will be presented for purposes of comparison. The case example data are shown in Table 2.2.10 Transferring nonnPL 566 technical assistance costs to the Federal side of the scales for the data shown in Table 2.2 gives a ratio of Federal, 347., and local, 6676.11 Compared to the national data, 60:40, the Michigan example project at 34:66 represents a much better degree of local participation, exceeding even the 50:50 ratio envisioned in the original PL 566 legislation of 1954, with all ratios expressing the percentage of Federal cost first.12 However, a second ratio can be computed for the example project, and it is almost an exact reversal from the first at Federal, 647., and local, 367.. This ratio counts ACP payments and planning as Federal costs, whereas the SCS ratios do not consider ACP payments because they are non—PL 566 costs, and count planning costs as non-project \ loSCS watershed work plans show land treatment and structural costs onJ-Y- For this project SCS estimates of capital costs for land treat- I'I'IEnt ($1,330,310, for the 73 square mile watershed area) and the bassociated cost" investment ($1,310,321, for the 14 square mile project- enefited area) are coincidentally- about equal, disregarding diminution of the latter for partial and delayed completion of the investment. as llln Table 2.2, PL 566 funds total $731,587, and other, $1,504,655, ‘ Show in the work plan table. Transferring nonsPL 566 Federal funds Feée. , PL 46 and Forest Service, or ,"going program" funds) to the. t e:L‘alside; Federal $770,067, and other, $1,466,175. From this $77 Percentages are as follows: Federal, 34.47. (0.344 = 0.067 /.$2,236,242), and local, 65.6%. C0 lzusnn, scs, D.A. Williams (scs Administrator), letter to state Asn§ervationists (t0p SCS officers, in each state), subject: "Federal SlStance in Watershed Projects," SCS Advisory WS’ZS, November 18, 1965. eem.mmew II I“ L’\ ' moo.o~om ”1.09. m0~ dem.eomw A.fiE .Um 0H okHCO m00fi>h®m COHUQHHQUWEH COfiUODNUMCOU QQHW UQUfiMQCCQ uUUfiOhQV OUEOU HmhauUDhum m504ufiu3QEOU Odu laud. U\n~ CH UWJCDOIO EUZDH H3240 3123“ 00m lo ‘l- IUI‘II'I Ill tilll SCI—N 1H0); Cu C3323 Nutmenhruufluv when umoo M0 oan II.-.’ D ‘I lll..lu.|"l I- I'll 'l :5!4:U#Z sjflzmflgdaz XQQHO Hfifiz No SDSQHQ :ehoz umunoo UUPWOUN “AU mun-W2 lfleOWH-Hsfl tho \fllde-SC-me «Noun 21 .uaoo N was H meromou .oonmnoums noos\o~m.oow pmo%\ownaN noos\ooe.~mo mm~.oa~.~o xoouo HHHS mo nocmum :uuoz Mom coauMuaoasooo .mom .uom umouom ..umwm Hooaanoos I I I I oom.- omm.meao mom ..oooo Hooeoeooa I I I I omo.mmH I I I I moofiuomue cowum>nomaoo nonuo A.oommm ma hacov omN.mmmw I I I I owmcfimuo A.fla .vm Mm .moum nonwuoumzv oumou ucoaumouu mama Hmm.oam.am I I I I I I I I Hoooooom oem.mn I I I I I I I I mcfimuo humuanwuu auwMIuoucH mwm.me I I I I I I I I AoHSumme o mooos weapon pomv coaumumeoue mama summlno oom.aeo.am I I I I I I I I oaeo aeoeIoo A.Ha .vm ma .waoo mono oouwmocon uoonouev mumoo oouofioommm Nmm.moom mam.mo~o NON.mooo Hoooooom ome.moa omm.mea I I‘I I ooooeoooo eo .aoo .oooo moo.moe I I I I mom.moe oooeeeoo ooeooHHooooH moo.o~ow Hom.mmao oom.aomo ooeooaeooeoo A.HE .vm ma .maao mono omufimoaon uoohouov mumoo Houauosuum macauousmaoo oHu mocsm Honuo moose comtum umoo mo oaHM Imu o\m ca omuaaou swam xnoz as asonm muomoumo commas .853; uses 32 no Sees fieoz .ooooo noose eo some some to tease oNoN QHQWH (PL 566 overhead) or could be computed, 5 into account, but ir least more indicati l4 ratios. To be so proportion of costs (1) the unavailabil under the Michigan Payments, $2 ,500 pa $10I000 per person funds due to Federa 22 (PL 566 overhead) costs.13 With other assumptions, different ratios could be computed, such as if limitations on ACP payments were taken into account, but in the author's view the ratio given here is at least more indicative of Federal cost sharing than the agency type ratios.14 To be sure, several limitations might reduce the Federal proportion of costs of PL 566 projects when ACP payments are counted: (1) the unavailability of ACP payments for established legal drains under the Michigan Drain Code; (2) the limitation on maximum ACP payments, $2,500 per person per year for non-pooling practices, and $10,000 per person per year for pooling practices;15 (3) lack of ACP funds due to Federal budget or program restrictions; and (4) farmers' willingness to apply conservation practices without benefit of ACP cost—sharing assistance. 0n the other hand, while the usual maximum ACP cost-sharing rate is 50%, it is interesting to note in the PL 566 context that flood-control type conservation practices have a higher rate of assistance (such as 807. for practice C-7) than tile drainage (330-507. in Michigan, practice C-lO). Rates higher than 507. may be used 13WPH (1961 ed.), sec. 1131.4 states that planning costs are not to be counted in the benefit cost ratio. b 14P1anning costs are difficult to estimate on a by-project basis, 13'; they may be in the $20,000 to $50,000 range. Using the ACP rates for 65 (Source, see note 15), and the list of practices in the example §r°JQCt's work plan table 1, ACP payments were estimated as $633,195. pitimated Federal costs are $1,453,262 ($731,587, as shown in Table 2.2; 88:8 $38,480, non-PL 566 technical assistance costs; plus $633,195, Fedimated ACP payments; plus $50,000, estimated planning costs). The in eral percentage would be 63.67. ($1,453,262/$2,286,242, the denominator lclmiing the total shown in Table 2.2, $2,236,242, as shown in the work an, plus $50,000, estimated planning costs). 15 Ha USDA, ASCS, State Office in Michigan, Airicultural Conservation Wforl965, Michigan (East Lansing, Michigan: ASCS, November 4). sec. 3, item H. ' 23 for practices "which have long lasting conservation benefits," and/or if an "increased rate of cost-sharing is essential to introduce a greatly needed new conservation practice."l6 EDA Grants Special Federal grants may be available for small watershed projects (1944 Flood Control Act and PL 566 projects) in areas that qualify for assistance under the Public Works and Economic Development Act of 1965 (PL 89-136, sec. 101). The minimum local rate of partici- pation for costs subject to EDA underwriting is set at 20%, and the converse is that the maximum rate of Federal assistance would be 80%.17 In practice this would increase the rate for all SCS project purposes, except flood prevention which is already at the 100% Federal rate. An Economic Development Administration grant was requested for the East Branch of Sturgeon River project, Dickinson County, Upper Peninsula, Counting costs subject MiChigan (SCS work plan dated February 1966). mderwriting with an Economic Development Administration grant (i.e. , CODStruction, installation service and administration of contract costs, and excluding land costs), the Federal government would have paid 45.4%, witl'lout the grant, and 75.9% with the grant, approaching the 80% maximum allowed . \ l6 Ibid. and sec. 10. 17 WPH (1967 ed., revision dated June 1968), Appendix 13: of umisrstanding between the Economic Development Administration-- Partment of Commerce and the Soil Conservation Service--Department of "Memorandum De rleulture . " Under PL 56 determined by: functional and p costs to these c< Reservoirs: facilities method purposes.19 Spec pose) are first d- according to the 1 alternative justij COSt remaining be: t‘10 numbers, eithe that Purpose with Specific costs dedI method is similar , 24 gost Allocation Under PL 566, the financial question of who pays for What is determined by: (1) Federal-local cost-sharing ratios for various functional and purpose cost categories; (2) the assignment of project costs to these cost categories; and (3) joint-cost allocation rules. 18 Reservoirs: For ease of explanation, SCS prefers the use of facilities method for allocating joint reservoir costs to project purposes.19 Specific costs (for items used exclusively for one pur- pose) are first deducted; then remaining (joint) costs are allocated according to the physical capacity assigned to each purpose. For the alternative justifiable expenditure method (also called the specific- COSt remaining benefit method), the allocation base is the lesser of tIWC) ntimbers, either the benefit of a purpose, or the cost of serving that purpose with an alternative, single-purpose structure, each with SPeeific costs deducted.20 The separable-cost remaining benefit method is similar, except that separable rather than specific costs are used in the initial deduction.21 \ l8See PL 566, sec. 3. Cost sharing rules are contained in: (1) WPH (1967 ed., relevant section amended as of Jan. 1969); (2) Economics W, ch. 10 (some pages amended as of Feb. 1968). 19 Capacity serving more than one purpose is divided equally. Sedi- Zlentation capacity is assigned to flood prevention only if downstream ' edillrlentation problems are alleviated, otherwise it is ignored. 20 C For example, if the benefits for irrigation are $40,000; direct $gstS-p $5,000; and the cost of a single-purpose irrigation structure, 5:000, the allocation base for irrigation would be $20,000. 21 ' Pu Se arable—costs are defined as. the difference in cost for a multiple a(lipose structure with and without the purpose in question when it is ed last. Specificficosts are those used exclusively for one purpose. W improvement cost: been reduced in 1 division method 1 the older method is more complex. ‘ Contractor Pavmen SCS develops P1566 project's I and the local 5pm 2'3» it may be use functional items c TVO basic cos one, for land cost Ia11d values; the s ICS in~state Plann- esr'. . lusting all othI 25 Channel improvement: SCS procedures for allocating joint channel improvement costs, usually between flood prevention and drainage, have been reduced in number and changed (in 1968 and 1969). The equal division method requires little explanation, and it is different than the older method shown in Table 2.3. The modified relative area method is more complex. 22 Contractor Payment Ratios SCS develops contractor payment ratios for inclusion in the signed PL 566 project's work plan agreement between the Federal government and the local sponsors. Before discussing their derivation in Table 2-3, it may be useful to describe the SCS development of the various, functional items of cost to be used. 'IVo basic cost estimates are used in PL 566 project work plans: one, for land costs, is based on SCS or local sponsor appraisal of land values; the second, for construction costs, is developed by the SCS tin-state Planning Party engineers, and it serves as a basis for estimating all other non-land costs for installing the structural mlea‘Sures. The SCS engineers use bid abstracts for other projects in the area, local material prices, and other costing resources common to \ d 22The modified relative, area method is used only if flooding and IVal-“age are joint problems on part of the project-benefited area of e Watershed. Wet land is defined as that portion of the area served 11: the channel already having or requiring on-rfarm drainage; and the “We: portion is. the remainder of this area. The preportion of the annel cost allocation to flood prevention is determined by the ratio: ’Sacreggefof non-wet land served by the channel) (total acreage served by the channel) ' R emajfling costs (i.e., those not allocated to flood prevention using t ehe area rule), are allocated to flood prevention and drainage on an Qual basis. the engineering p1 computed with rule cost estimate, as in reaches 1 and for 1961). COHStruction Basic engin Con tiflgeHCy GODS truct InStallation 5 Engineering , sUbtotal other, 52 o 26 tflne engineering profession. Other nonnland functional cost items are ccunputed with ruleeof—thumb percentages from the basic construction ccust estimate, as shown in the following data for the channel improvement 111 :reaches 1 and 2 of the North Branch of Mill Creek watershed (prices for 1961). Construction cost Basic engineering estimate Contingency allowance, 15% of $434,516 Construction cost, subtotal Installation service cost Engineering, 10% of $499,693 Subtotal Other, 5% of $549,669 Akdministration of contract cost 5% of $499,693 (construction) Land costs (easements and rights of way) frotal installation cost for this structure, excludes both planning costs and land treatment costs These costs paid from Federal (F), local (L) or shared (3) funds, as Show! in the last column. $434,516 65,177 $499,693 49,969 (549,669) 27,483 24,984 93,780 $695,909 The allocation process will be discussed in several steps: Step 1: Since benefits were used as an initial allocation base to divide the cost for all structures (including the flood water S 27 .cmam xuos cw uouum .NomH mumaunom vmumv .aoHuMucmadoov mam swam xuoa .mom .«omb "woumanm Momma omaw memm.qmmw mew omm.mqmw .ssonm mmwmuaoouoa do momma cowumooaaonm voumooaam mumoo owwaamum mmoq eeo.mm~w www.meam o~m.~mam uuuou umuueuuu Heuoe euuuuoaa< NOOH Nmm qu mmmdfimhv .mmmn GOHumUOHHm UGNHHDU mHQMUHHmm< mom.mmew -- -- eeo.mmmw omn.omw oem.~mm mem.ooem woe.mew ame.mmmm Huuouesm ow~.mm om~.mm -- mee.mm men.mm --- “Ho.em NHo.em --- uses ewm.e~ mam.oam -- mam.oa mam.oem --- Ham.eH Hmm.eam --- uueuueou no .au< mme.nm -- mme.a~ mme.HH -- mme.HH omw.ma -- omm.nH eutuo mem.me -- oes.mee awH.H~ -- NmH.HNm NmN.mN -- Nmm.m~ meauuuuamem moofi>umm .HumcH mme.mmem Aeoaue uumv cem.HH~w Asoaue uumv mmn.mm~w -- m-.~mmm coauuuuuueou mmeHm-fiu N<.N¢ mam coaufimurmua @OOHH N©.Nm “mumou HOGGmSU mo GOfiHMUOHH< No.o0H Ne.~e No.5m omen eoeuuuoaae Huueueu weeuaeuum mom.mmem eeo.mm~w new.ooem Huuou .uuou Hueeeeu euuuuoaaa mmmhqual mmmaqual .uuduum wGHvaumH HmumspfiOOHm mmmA eeN.oemm eeo.mm~m oo~.memm uuuou euueuoaaa No.00H NH.mm Nm.ee museueue aoue uuue eoeuuuoaaa Hence uueuo eon um Heuoa uueuo gem am Hence uueuo com um m moo m -. malnu-n- =m>uum- oo some uuou ”HHHHNHNHHHHHHHHHHW u H uoe mumou m mcfimua mumou cow“ b ah nMmuuuHnuunuunnunuuuuuuuuunuuuununnnnunnuuHHH. . . . m as eueuuuuuz euuuo Hafiz mo aueeum euuoz .N ecu H uueueum Mom muemameoumau Hmucmgo .eowumoomww “new m N Ha 28 retarding structure) 23 in reaches. 1 and 2, the percentages 64.9%, flood prevention, and 35.1%, drainage, were applied to this cost $840 .264) . 24 Step 2: After deducting the flood retarding structure's cost, a second allocation base was derived for the channel improvement costs only, 57.6% flood prevention, and 42.4% drainage and applied to all functional cost items. In combination with the by-purpose, by- function cost—sharing ratios, this allocation is sufficient to separate PL 566 and other costs assigned to flood prevention, but an additional allocation is needed to separate drainage construction costs. Step 3: At the time the work plan was originally completed (in 1962) , SCS regulations required that other (non-PL 566) funds bear 55% Of all costs allocated to drainage. Already allocated functional cost The result is the items (from step 2) for drainage were then deducted. diVision of drainage construction costs between PL 566 and other funds. Construction costs are then summed, for PL 566 and other Step 4: f‘mdB, and the result determines the contractor paying ratio, 77.6% Faderal and 22.4% local for this work. While the contractor paying ratio \ 23The flood retarding structure costs (for reaches 1 and 2) were all(mated and divided in a less elaborate manner, simply using PL 566 SOStwharing ratios for functional cost items, for all of the costs ere allocated to flood prevention. The channel improvement costs for reach 3 were allocated using a different base than was used for reaches and 2.. ’ 24 A See ch. 3. For this project, FPB = FWDRB + 1/2 MILUB + 1/2 LUCB: "“3 (drainage only) = 1/2 MILUB + 1/2 LUCB. 29 is based on estimated costs, it becomes the basis of sharing actual costs; as such it becomes part of the legal, work plan agreement between the Federal government and the local sponsors. Flood Prevgntion Dominance The 1967 House Agricultural Committee policy statement, requiring that flood prevention be the dominant purpose of all PL 566 projects the Committee approves, is less restrictive than it seems.25 First, because flood prevention receives a higher degree of Federal underwriting than other purposes, its dominance is consequently more certain when relative PL 566 cost for structures is used as a 26 criterion. Second, not all PL 566 projects must be approved by the House Agricultural Committee. The act requires that Congressional approval must be obtained if the Federal contribution to construction cost exceeds $250,000, or if overall capacity of any single reservoir exceeds 2500 acre feet. Below these limits, stated in PL 566 itself, con- struction fund allocation may be administratively approved within SCS. If any one reservoir has a capacity in excess of 4,000 acre feet, construction must be approved by the public works rather than the agricultural committees of both Houses of Congress. (The maximum size 25See USDA, scs, D.A. Williams, scs Administrator, Watershed lhmmrandum.86, subject: "Flood Prevention in Watershed Projects," dated September 28, 1967. Attached is a letter from K.R. Poage, Chairman of the House Agricultural Committee, to Speaker of the House, JO n W. McCormack, dated July 31, 1967. 26On this basis, the project cited in Table 2.3, would be 76% flood Prevention compared to 64% when all structural costs (PL 566 and other) are considered, and 65% in terms of benefits. is 25,000 acre f original minim SCS memorandum! I apply ‘0 PrOjecu because 0f their Public Works Com: In the slime program, issued b shed developments land into product and emphasizes mu restriction on fl- statement elimina or depressed rural tal assistance 01' Iron SurPlus crop aBricultural uses 30 is 25,000 acre feet at present, as stated in PL 566, although the original maximum'was only 5,000 acre feet). According to the cited SCS memorandum, the House Agricultural Committee restriction does not apply to projects which can be administratively approved within SCS because of their small size and cost, or which must be approved by the Public Works Committee because of their large reservoirs.27 In the summer of 1967 came another restriction on the PL 566 program, issued by the Department of Agriculture.28 It restricts water— shed developments that are primarily intended to either bring new land into production or to increase output of crops already in surplus; and emphasizes multiple purpose projects, significantly omitting any restriction on floodwater damage reduction benefits (FWDRB). Also, this statement eliminates restrictions of any sort on projects in low-income or depressed rural areas, and stresses coordination of various Departmen- tal assistance or subsidy programs to speed the conversion of land uses from surplus crop production to non-surplus crop production or non- agricultural uses. 27 PL 566, sec. 2, item 2. For example, the Mill Creek project, mentioned in the preceding footnote, had an overall structural capital cost (SCK) of $905,932. Of this, the Federal contribution was $607,207. Deducting installation service costs from the Federal cost, $501,304 remains as the PL 566 contribution to construction costs; since this exceeds $250,000, the project could not be administratively approved within SCS and required approval by congressional committees. Approval was obtained from the agricultural committees, for the reservoir had a capacity of 1670 acre feet which is less than the 4000 acre feet capacity Signaling Public Uprks Committees approval. 28USDA, scs, scs Administrator, D. A. Williams, Watershed Memo- randum 84, Supplement 1, subject: "Surplus Crop Production," dated. August 25, 1967. Attached is a letter to the Secretary of Agriculture (Orville Freeman), from.John A. Baker, Assistant Secretary of Agriculture for Rural Development and Conservation, dated July 18, 1967. 31 An Overview Taking the Department's and the House Agriculture Committee's statements together, it would appear that flood prevention is to be the clearly dominant purpose of PL 566 projects in terms of PL 566 costs for structures, and this is not necessarily in conflict with the Depart- ment's emphasis on multiple purposes in terms of benefits, because the higher degree of Federal underwriting for flood prevention means that it can more easily dominate among the purposes in terms of costs. Significantly, if one is looking for a sense of complementarity with the Committee's statement, there is no restriction on FWDRB (reduced losses from flooding). However, the restriction that project benefits can not result primarily from bringing previously uncrOpped land into crop production has been extended to include the flood prevention portion. Yet, the Department's policy statement makes no mention of non-FWDRB flood prevention benefits associated with more intensive land use of already cropped land (MILUB-FPB). However, MILUB—irrigation and MILUB— drainage must be associated with furthering "efficient use of water and related land resources," rather than with increasing crop production per se. Conservation and PL 566 In a typical PL 566 project, conservation (land treatment) practices receive considerable emphasis. They are partly financed by ACP payments. Technical assistance for their application is provided by USDA personnel, from SCS (PL 46 funding) and the Forest Service. Accelerated application is funded from PL 566; this may be necessary if a project would place too heavy work loads on regular, in—county USDA personnel. Application in critical areas is apparently not well 32 understood and is to be distinguished from that in nonncritical areas. In this section the role of conservation districts, farmer agreements and SCS emphasis will also be considered. Critical Area Land Treatment For small watershed projects, Hugh Hammond Bennett, first SCS administrator and founder, indicated that it was first thought to be necessary to complete 100% of the critical, runoff-reducing and erosion-reducing land treatment on the watershed drainage surfaces above a planned flood-detention reservoir E£12£.t° initiating any construction on the structure. However, Bennett later indicated that this prerequisite could be reduced from 100% to 80%. Morgan incorrectly compares these percentages to the one mentioned in PL 566, 50%, which refers to the proportion of land above a detention structure that must be under agreement "to carry out recommended soil conservation measures and farm plans."29 29The quotation is from PL 566, sec. 4, item 5. Critical runoff and sediment producing areas must be distinguished from other watershed areas. Not all watersheds have these critical areas; they are absent in most flatland areas such as Michigan, and seem to be peculiar to certain geographical areas of the country with special soil, topography and rainfall combinations. The Southern Plains seems to be one such area. Reservoirs built in such areas ‘have been subject to extremely rapid and unexpected sedimentation fill- in, Owing to lack of erosion control in the watershed areas contributing runoff to the reservoirs. 33 Morgan drives his criticism home by adding; There was, and is, no statutory requirement that any land actually be treated. Local sponsors have a supposed obligaa tion to effectuate agreements for completing and maintaining these projects, but, since this responsibility does not have to be met, there is little ground for believing that it always will be met.30 However, SCS regulations indicate that SCS expects 75% of the critical—area land treatment to be completed concurrently with the construction of flood control structures on the main stream. Thus, 75%, not the 50% in PL 566 itself, may be compared with Bennett's 100% and 80% figures, as the current, expected measure of runoff and erosion control in critical areas above flood detention structures. As a matter of—fact, these SCS regulations negate the PL 566 legislative requirement: (1) if_erosion and other problems in the area above the flood detention structure would not adversely affect the structure (in terms of design, cost, and operation-maintenance); ppd_(2)_if farmers in the specified area are soil conservation district cooperators.31 30Comparison and quotation from Morgan, pp. 188-189; also, see pp. 178, 299 and 300. 31See WPH, 1967 ed., sec. 104.03; sec. 1110, 1961 ed., is virtually identical. Farmer agreements with soil conservation districts are made on three different levels. In stage I, the farmer merely signs an agreement with the district, indicating his interest in following conservation-oriented practices. This,mekes him eligible for assistance from the district, since he is then a district cooperator; that is, the SCS technician can then aid him in installing various conservation practices. Stage II means that the farmer has allowed the SCS technicians to examine his land and perform a conservation survey upon which a map of land capabilities could be prepared. In stage III the farmer agrees to accept and execute the basic conservation plan. TheSe comments are based on Morgan's discussion, pp. 156-168. 34 Critical erosion and runoff areas in the watershed drainage area above floodndetention-structure sites are not a problem in most relatively flat areas, such as Michigan. 32 In regions where these critical areas are a problem, farmers can apply to their county ASCS office for Federally-funded ACP payments to cover up to 80% of the cost of such critical-area conservation practices.33 In addition, PL 566 funds are available to supplement ACP and other "going program" funds "for planning and application of land treatment measures" in critical areas and elsewhere, if there is a lack of funds under these other national programs (underline added).34 However, PL 566 itself stipulates that the overall Federal technical and cost-sharing assistance under PL 566 shall not exceed that available under other national programs.35 Therefore, either under ACP or PL 566, it appears that the Federal government will pay up to about 80% of the cost of installing critical- area treatment measures, if classified as land treatment practices in the project work plan. Different rates of assistance apply for other types of land treatment practices in non—critical areas of the watershed; 50% may be a roughly typical rate under the ACP program, but specific rates are both lower and higher. 32 SCS work plans for PL 566 projects in Michigan usually indicate that field investigation revealed that sedimentation damage and problems area not serious. 33Assuming that the 80% Federal cost-share for ACP practice C—7 (Structures for erosion control) forIMichigan for 1965 is typical. Source: USDA, ASCS, State Office in Michigan, Agricultural Conservation QEEOBFam Handbook for 1965 (East Lansing, Michigan, ASCS, November 1964). National ACP practice Cv6 (storage dams for erosion control) is not listed in the ACP Handbook for Michigan. 3433;, sec. 104.04. 35Compare Ibid. and PL 566, sec. 3, item 4. Alternatively, ‘ classified essential than as land treatise this case, the Peder cost, excluding lan- operation-maintenan SpOIlSOIS . Son-Critical Area L Adjoining the the critical-area 1 installation of the regulation concerni Which is usually in determine" that a 1” by the project strL prevention is omitt district or equival i’hich a re Prepared 35 Alternatively, critical-area conservation practices may be classified essentially as flood prevention structural.measures, rather than as land treatment measures in the PL 566 work plan dichotomy. In this case, the Federal government pays for 100% of the installation cost, excluding land, administration of contract (until 1969), and operation-maintenance costs, all of which are paid by the local sponsors. Non-Critical Area Land Treatment Adjoining the SCS administrative regulation requiring that 75% of the critical-area land treatment be completed concurrently with the installation of the mainstream structural measures is a counterpart regulation concerning land treatment in the project—benefited area which is usually in the valley lowlands. It states that SCS "will determine" that a high proportion of the farmers on the land benefited by the project structural measures for irrigation or drainpge—-flood prevention is omitted--will agree with the local soil conservation district or equivalent, not SCS, to "develop" farm conservation plans, which are prepared by the in-county, local SCS technician.36 The Role of Conservation Districts Soil conservation or other similar districts now cover virtually all of the United States, with the exception of a few areas that either lack interest in agriculture or oppose the SCS approach. 36 ‘WPH, sec. 104.036, item 3. The agreement by the farmer with the district to develop the SCS conservation farm plan means that the farmer would be in stage III of progressive planning which is discussed in footnote 31. 36 In the PL 566 context, districts lack taxing power and therefore can not be financially responsible for the local share of project costs. They may be effective in promoting projects locally, and in making various local contacts necessary in this effort.37 Farmer agreements with districts have already been discussed.38 SCS Emphasis on Land Treatment Since the cited PL 566 legislative and SCS administrative regu— lations are specified in terms of farmer-local district agreements, rather than in terms of legally binding contracts, one could argue that these agreements, not land treatment and conservation per se, are by subterfuge the condition upon which Federal assistance is provided to local people. On the other hand, SCS has traditionally prided itself as being a, if not the, leading Federal agency interested in promoting and plan- ning soil and water conservation. SCS regulations accentuate land treatment by discussing it in the Opening paragraph of the section on watershed planning.39 Land treatment has been emphasized in directives from the SCS Administrator, who, in one extensive directive on the subject, states: All Service employees must recognize the necessity for adequate land treatment in watershed projects. Needed actions must be taken to insure that each watershed project when completed will 371Morgan, ch. 12, especially pp. 338v342. 38See footnote 30. 39WPH, sec. 104.00. 37 be a "showcase" of sound land use and treatment. The quantity and quality of conservation treatment on the land should be an identifying mark of any completed watershed project.40 An advisory notice from the SCS Administrator suggests that on- site inspections of watersheds usually reveal an adequate amount of land treatment. Subsequently, he comments that SCS policy has and will continue to stress "that land treatment is the keystone of watershed development."41 Planning and Coordination PL 566 calls for special consideration of surveys and plans of the Department of the Interior with respect to the conservation and develop— ment of fish and wildlife resources.42 A PL 566 project may be affected when consultation so requires in relation to legislative authorities of other departments, as stipulated in Presidential Executive Orders.43 The Corps, generally, has responsibility for larger (downstream) watersheds with drainage areas of 250,000 acres or more, and SCS, for smaller (upstream) watershed areas. In addition, for urban areas, 40USDA, SCS, SCS Administrator, D. A. Williams, Watershed Memo- randum-70, subject: "Watershed Land Treatment," dated November 5, 1964, p. l. 41USDA, SCS, SCS Administrator D. A. Williams, Advisory Notice We748, subject: "Watershed Protection (PL 566)--Land Treatment Measures in watershed WOrk_P1ans," dated September 28, 1962. 42PL 566, as amended, sec. 12. 4‘3The small watershed authorities are; the Flood Control Act of 1944 (58 Stat. 887), as amended, for the eleven river basins; nationally, including the 50 states, Puerto Rico and the Virgin Islands, PL 566 of 1954 (68 Stat. 666) as amended. The reclamation acts are;' the Reclamation Act of 1902, as supple— nnnted and amended (43 U.S.C. 391); the Small Reclamations Project Act of 1956, as amended (43 U.S.C. 422a—k). The Flood Control Acts of 1917 (39 Stat. 948), 1928 (45 Stat. 534), and 1936 (49 Stat. 1570), all as amended. 38 Corps has responsibility, if major damage would occur ($2,000,000 or more); SCS, if ndnor damage would occur ($750,000 or less); and responsibility is subject to negotiations and further guidelines, if intermediate damage would occur ($750,000 to $2,000,000). Urban damage is decided on the basis of a flood sufficient to inundate "substantially the entire flood plain.“4 Water resource development activities are divided among several federal agencies. Several Presidential Commissions have proposed a single agency, but perhaps the best that can be expected is inter— agency coordination. The present Water Resources Council is apparently the first to receive congressional sanction, although inter-agency groups date to 1939. While this tOpic has a rather interesting history, it will not be considered further here. The Tennessee Valley Authority Act of 1933, as amended (16 U.S.C. 831, et seg.). “See USDA, $03, 303 Administrator, D. A. Williams, Watershed Memo- randum 75, subject: "Agreement with Corps of Engineers with Respect to Flood Protection by Engineering Works," dated December 14, 1965. This memorandum transmitted the Corps-SCS agreement dated September 23, 1965. This agreement was a condition for favorable action by the Senate Committee on Agriculture and Forestry on Public Law 89- 337 (approved November 8, 1965), amending PL 566, which.increased the limitation on flood detention capacity for PL 566 project reservoirs erm.5,0QQ to 12,500 acre feet. Overall capacity for PL 566 reservoirs is 25,000 acre feet, including capacity allocated to all purposes. These limitations refer to single reservoirs, and a project may involve several reservoirs. 39 Flood Versus DrainagefPrpblems Floods may connote disaster to many people, and their control takes on the meaning of such terms as national interest, national defense and national welfare.45 Disastrous flood losses that attract national news coverage usually occur in large river valleys, but less dramatic losses in smaller, upstream areas may account for over half of the annual flood damages for the United States, according to SCS estimates.46 There is the impression that the downstream program is intended to control large, disaster-type floods; the upstream program, frequent floods. Yet, neither program affords complete protection.47 Both the Corps of Engineers and SCS employ low-probability design floods, and justify protection works on the basis of reduction in mathematically expected annual damages, to which low-probability floods 45The preamble to PL 566 states in part: "erosion, floodwater and sediment damages in the watersheds of the rivers and streams of the United States, causing loss of life and damage to property, constitutes a menace to the national welfare." 46Erwin C. Ford, Woody L. Cowan and H. N. Holtan, "Floods--and a Program to Alleviate Them," in USDA, Water: The Yearbook of Agriculture, 1955, (Washington, D.C.: USGPO, 1955), pp. 171-176. The authors use data for 1952, for floodwater and sediment damage, of which the upstream portion is 56% for the United States. Upstream damages, $557 million, were estimated by SCS from studies of 77 watersheds covering 52% of the continental United States, over a 15 year period. Downstream damages were estimated as $500 million. LeOpold and Maddock criticized an earlier, 1945 SCS ratio, 75% upstream and 25% downstream, as being based on a hypothetical watershed, some assumptions and extrapolation. See Luna B. Leopold and Thomas Maddock, Jr., The Flood Control Controversy (New York: The Ronald Press, 1954), pp. 186al88. 47Leopold and.Maddock, p. 239. 4O contribute very little. Highrprobability, low—damage floods contribute most to expected annual damages for a given location, perhaps 80% for lO-year and more frequent, smaller magnitude floods.48 However, Leopold and Maddock argue that there is a conflict in goals: projects are quite successful in promoting land development, but only partially successful in meeting loss reduction objectives, for protection is never complete. Some protection spurs floodplain deve10pment which becomes the basis for demands for more protection. They argue that zoning and other non-structural methods for avoiding losses are seldom proposed, because such methods conflict with local interest in land development and real estate value promotion.49 Some 10 years later a Presidential Task Force stated: Studies of flood plain use show that some flood plain encroachment is undertaken in ignorance of the hazard, that some occurs in anticipation of further Federal protection, and that some takes place because it is profitable for private owners even though it imposes heavy burdens on society. Large numbers of soundly conceived, economically justified flood projects have been built. As a result, vast flood damages have been prevented. However, vital actions needed to complement the structural protection effort have been absent. In consequence, the Nation faces continuation of a dismal cycle of losses, partial protection, further induced (through submarginal) development, and more unnecessary losses.50 48This refers approximately to the area to the right of the lO-year frequency line in Figure 3.1. For Figure 3.1, damages for the l, 2, 5 and 10 year floods may be obtained by summing the damages shown in the last column of Table 3.4: $36,288/$45,410 = 0.799 or about 80%. This represents the rectangular-area method of approximating the area under the SCS damage frequency curve. 49Leopold and Maddock, pp. 239-240. 50U.S., Office of the President, The Task Force on Federal Flood Control Policy, A Unified National Program for Managing Flood Losses, House Document No. 465, 89thCong., 2d Sess. (Washington, D.C.: USGPO, 1966), pp. 11*12. 41 Leopold and Maddock (who are hydraulic engineers, not economists) point out that flood protection projects were originally promoted on the basis of disaster relief, with many of the flood control acts following floods of major proportions. To emphasize the point, they indicate that the terms flood prevention and flood control are misnomers for flood protection which is never meant to imply complete protection. They propose that the program be stripped of the implications of disaster- relief benefits. Concentrating on the idea that flood protection projects have become relatively highly subsidized forms of land development, they outline the successive retreat in the requirements for local financial participation, going from the 1917 to 1938 Flood Control Acts, and propose local participation in proportion to benefits. They touch on the idea of comparative development advantages and cost for different areas.51 If flood protection is motivated and supported as a means of Federal-paid land development for local beneficiaries, is it any different than drainage or irrigation in an agricultural setting? 'r w.—vv—v— 51Leopold and Maddock, pp. 144 and 2409244. The study of comparative costs and advantages is different in purpose and viewpoint than that of project evaluation and justification. The former may require significant changes in data and procedural assumptions in agency evaluations. See Vernon W. Ruttan, The Economic Demand for Irrigated Acreage (Baltimore: Johns Hapkins Press for Resources for the Future, 1965), pp. 85-88. 42 This question is relevant to the discussion of SCS procedures for evaluating flood and drainage problems, for disparate Federal cost-sharing, planning approval (USDA) and construction approval (congressional) rules necessitate separating flood prevention and drainage benefits and costs. Although the whole matter could be left as a question of policy, some conceptualizing may be useful. In particular the nature of loss and flood-hazard effects on farm managers and land use (cropping patterns) are of interest. FWDRB are computed by SCS using the simplifying assumption of homogeneous cropping patterns in economic reaches, which are sub-parts of the project—benefited area. Yet, for the North Branch of Mill Creek watershed (the example in chapter 3), it appears that SCS's sampled farms with a higher prOportion of land in woods, idle and permanent pasture uses, and less land in crop uses were located nearer the river, although the author could not precisely locate the sampled farms on the watershed's map. If this is so, it indicates that the SCS simplifying assumption of homogeneous land use and cropping patterns disguises farm manager perception of flood hazards and consequent loss-avoidance reactions. Existence of uncropped land nearer the river is consistent with the hydrologists' view of the manner in which rivers develop and use floodplains to handle overbank flows.52 T—v T——v--w ‘ .1—T_ 52Leopold and.Maddock, ch. 2. Channels carry normal flows, and floodplains c0pe with occasional excess flow. Natural floodplain heights are determined by reasonably frequent, not rare flows. Hoyt and Langbein studied overbank flow at 140 locations in 36 states and found that, on the average, minimum damage stage coincided with the degree of overhank flow that is equaled or exceeded every two years. 43 Land use near the river is relevant in estimating FWDRB, because of dependence on frequently—flooded land, even though an entire reach may be inundated by say 25 or 50 or 100 year floods. That is, only the land near the river is flooded often enough to affect management behavior in an expected sense. Furthermore, the assumptions of the SCS model base FWDRB largely on mid-growing-season flood losses, when values subject to loss are highest, but if these occurred as envisioned in the model, one would expect loss avoidance reaction by farmers. Rather, evidence for Midhigan suggests that spring flood losses are more likely; their regularity could prompt loss-avoidance reactions in the form of late planting. Because FWDRB are based largely on without-project damages for relatively frequent floods (say 10 year and smaller floods),53 it is relevant to ask if these floods are perceived as being any more or less subjectively certain than drainage or irrigation problems. Definitionally, flooding implies river overflow, whereas impaired drainage relates to high water tables, although in the SCS "abnormal rainfall" construct, stream overflow is not the only cause of flood problems.54 However, this construct does not appear to be used in 53See footnote 48. 54See WPH, sec. 105.00; To differentiate flood prevention from drainage on flat lands, flood prevention is any undertaking for the conveyance, control and . disposal of surface water caused by abnormally high direct precipitation stream overflow, or floods aggravated by or due to wind or tidal effects. 44 SCS evaluations of FWDRB. Crop growth,.management decisions and farm income can be affected by excess moisture in the root zone, that is root zone flooding, regardless of whether the water is from river overflow, abnormal precipitation, high water tables or other causes. The distinction has been made important in terms of Federal cost- sharing for PL 566 projects, although the rationale for this remains unclear. As indicated in chapter 3, FWDRB are only a portion of the project— credited farm income, conceived as loss reductions rather than as gains. Enhancement benefits (EB) and the included farm income are conceived as gains. FWDRB are computed by taking into account the expected annual extents of flooding, with and without the project, on an assumed flood- free situation; of course, the watershed is not flood-free. Similarly, "drainage damage reduction benefits" or "irrigation damage reduction benefits" could be computed as analogs to FWDRB (floodwater damage reduction benefits), assuming well-drained or adequately-irrigated conditions, and taking into account the project effects or loss reductions. Flood protection, drainage and irrigation are different ways of achieving increased farm income and crop production. The policy-based preferential cost-sharing treatment of flood protection is no assurance that it is the least costly or.most effective way of achieving the objectives of increased farm income and output. This preferential treatment does of course work to the advantage of land owners whose water problems can be classified as flood rather than drainage or irrigation problems. 45 Summary The small watershed program was begun as part of the public works effort to increase income and employment during the Depression, under the conservation and demonstration projects work of the Soil Erosion Service (SES) in 1933. Emphasis on conservation has since distinguished it, for reasons of SCS interest and possibly of defense in the rivalry with other agencies. The small watershed program has been carried on under five authorities: the 1933 National Industrial Recovery Act, the 1935 Soil Conservation Service Establishing Act (PL 46), the 1936 Flood Control Act (especially the survey approval for the eleven flood prevention watersheds in the 1944 Flood Control Act), the 1953 Pilot Watershed Appropriations (in the Agricultural Appropriations Act), and finally PL 566 of 1954. Congressional public works committee approval was required under the Flood Control Acts (and still is for 1944 Flood Control Act projects), but is required only for larger PL 566 projects, those with reservoirs exceeding 4,000 acre feet and up to the legislated maximum of 25,000 acre feet. Otherwise, congressional agricultural committee approval is the rule, except for very small PL 566 projects, with reservoirs of less than 2,500 acre feet capacity or with Federal construction cost less than $250,000. These alternative project approval routes are critical, for the House Agriculture Committee will approve projects only if flood prevention is the unmistakeably dominant purpose in terms of Federal construction cost. This is less constrains ing than it appears at first glance, because there are other cost components, and because Federal cost sharing is higher for flood prevention. 46 Likewise, the Department of Agriculture's 1967 policy statement does not appear too constraining. Surely, all forms of land use Change ' benefits (LUCB), including the flood prevention sub-category as well as irrigation and drainage as in the past, are now restricted, but only in the sense that they can not dominate other categories of benefits. No restriction is placed on MILUB-FPB or on FWDRB, althougthILUB— ' rather than to drainage and MILUB-irrigation must be for "efficiency,' increase surplus crop production. None of these USDA restrictions apply to projects built in designated, low—income or high-unemployment areas. However, even with special Economic Development Administration (EDA) grants, the rate of Federal cost sharing for a predominantly non—flood prevention PL 566 project is likely to be lower in such an area than for a project else- where with flood prevention purposes dominating. Conservation is of primary interest to the Soil Conservation Service. In the context of small watershed projects the SCS, along with the Forest Service, plans land treatment practices. The Agricultural Stabilization and Conservation Service (ASCS) provides financial assistance for application of these practices. Conservation practices add up to an important investment for small watershed projects, possibly exceeding that for structural measures. Thus, ACP payments can con- stitute a significant Federal investment, but they are usually ignored in SCS computations of Federal—local cost—sharing ratios, possibly because they are administered by another agency, and possibly because SCS can not guarantee a specified percentage of ACP assistance. CHAPTER III THE SCS MODEL This chapter is concerned with SCS procedures for evaluating agricultural benefits, an emphasis based on that of the PL 566 program.1 These procedures are formulated into a model that is employed in the sensitivity analysis of chapters 5 and 6; they are further studied in chapters 5-7. This chapter is divided into several sections: (1) an overview; (2) FWDRB and hydrology; (3) FWDRB estimation: computational steps; (4) enhancement benefits estimation: computational steps; (5) project costs; (6) obtaining the benefit cost ratio; (7) data inputs, sources and assumptions; (8) net project effects; and (9) summary. An Overview The SCS model's investment criterion may simply be expressed as: B/C . (annual benefits) / (annual costs). This ratio must exceed 1:1 to justify the investment economically, including both Federal and local components, although other criteria.must also be met. Project struCtural coats for the major .mainstream project works, are counted fie 1See Appendix, Table 5, for benefit data. 47 48 in the denominator of this ratio, but associated costs for the complementary, on-farm and inter—farm works, are deducted from projects credited farm income in the numerator.2 In the PL 566nSCS context, the watershed is a complete surface drainage basin, usually without any additional upstream drainage area, and legislatively limited to 250,000 acres (about 400 square miles), although projects may be planned to adjoin one another. For the North Branch of Mill Creek project, which will be used as an example in this chapter, the water— shed includes 73 square miles, but only 14 square miles are contained in the smaller project benefited (or project benefit or interdependent) area. The project benefited area is in turn subdivided into more or less homogeneous areas, known as economic reaches, on the basis of economic variables, notably cropping patterns and type of agriculture. Several hydrological reaches are typically included in one economic reach. This presentation of the SCS model will depend extensively on previous work by the author.3 In systematizing the SCS procedures into 2The idea that benefits should exceed costs is expressed in PL 566, sec. 5, item 1, and traces to the Flood Control Act of 1936, the previous authority for the small watershed program, as discussed in ch. 2. Investment criteria are further discussed in abs. 4 and 7. Relating Federalnlocal cost sharing (discussed in chs 2) to the SCS investment criterion, it would appear that this criterion incorporates mixed budget constraints. Both.structural and associated capital costs include local and Federal components, assuming consideration of ACP payments in the latter (see cost data for 12 Muchigan PL 566 projects, Appendix, Table 1). 3John Vondruska, Estimating Small Watershed Project Benefits: A Camputer Systematization cf SCS Procedures (East Lansing, much. : Dept. of Agric. Econ., Michigan State Univ., Feb. 1969), Agric. Econ. Report 120. 49 .mGOAquaoo covooam use .voaHmHvIHHmB How ma suuaow can “moaned mmnsm men ou coaumooaam ucoquMfin m on coma newfia nuuaom m use .wouamaoo mum oaooca um: mo mHo>oH manage .vouoawH mum muoaumumaom uamooon uaoaowmcma nouns HousuaaofiuwmIGOHuco>oum vooam can .mao>ma maooafi aumm noofioumuauas onu mo usuao>oasoo vomoaov can Hosanna How muaaaumsfivm .muowuoacov umoo woumfioommHm90w>oua a AmDAsz vomaouo mummuao uom mouse Iaoo our mufimoaon uuoaoocmncm .N mooauavcoo vocawuplsauooa msaaooflm sass . .oaooua um muowufivaoo ouaouvlmauoom w moumlvooam .mEOUGH um umzoaaom mm woumaaoawo .ocon coauoapou vooam ou map oaooaa um: ca mwcmno on» ma moaned mMQZM .H .noaumaufim Ammoaloc Ho vumnonlouoNv moHMIvooam voabmmm so an o>aumaou Amumoo coauoavoua cw ommouoow no no no «sauna mmoum qu mmoa one ma nuance no common Hmnuao Home: ommamn .uuohoua onu On our oucoHoHMHu ommsmv u mmnzh .H mucouomaoo on museum“ add a“ owcoso HHouo>o can no wouwasoaoo mufimocom communes maoomfi use on nonmaaoamo muuocogaoo uawoamm muamocon uuoaoocunuo was mmnzh mm woumaaoamo muuonomaoo ufimoaom m.ponHHmaHm .moHuowoumu Hmaofiumusmaoo ufimocom mo ocaauao .H.m manna 50 a model, what is believed to be typical SCS practice has been selected for presentation here, although it should be realized that some procedures are alternatives to others.4 Agricultural Benefit Evaluation Alternatives To evaluate the benefits of the total package of investments and changes associated with a PL 566 agricultural project, one needs an estimate of the change in net farm income. Associated costs are deducted; then an adjustment is made to reflect the partial and delayed achievement of the with-project net income level, and the result becomes the numerator of the SCS benefit cost ratio. See column 3, table 3.1. However, in addition to distinct and separate cost allocations SCS usually categorizes agricultural benefits into several components, for congressional and administrative approval may depend on the importance of various kinds of benefits as indicated in chapter 2. Compared to the process of estimating overall net-income change benefits, different assumptions are employed if FWDRB (floodwater damage reduction benefits) are estimated separately. Hydrological data on flood occurrence and estimates of crop damage are needed. Initially, flood-free (zero hazard) conditions are assumed, then damages for the expected annual extent of flooding are computed for 4See USDA, SCS, Economics Guide (Hashington, D.C.: SCS, Mbrch 1964). USDA,-SCS,'Watershed Protection Handbook1 Part 1: Planning and ‘ngratio 3 (Washingto , D.C.: SCS,August 1967). USDA,SCS, National Engineerigngandbook, Section 4,,HydrologyJ Part I: Watershed Plannigg_ (WaShington, D.C.: August 1964), prepared by Victor Mockus. In comr- nnn with SCS practice, abbreviations will be used: ‘Ec0nomiCs Guide, lEEEJ and NEH—4, respectively. 51 both with and without project hydrological conditions. This method of separating FWDRB and enhancement benefits5 is shown in column 1 of Table 3.1. Alternatively, an analog to FWDRB may be computed, along with enhancement benefits, without using hydrological data, using the approach shown in column 2, Table 3.1. FWDRB are a partial estimate of the aggregate of net income changes associated with moving from the crop production levels without the project (high flood hazard and poor drainage conditions for most Michigan projects) to those with the project (for low flood hazard and artificially well-drained conditions). With reduced flood hazards-~flood protection is never 100% complete-- net returns in addition to FWDRB accrue to farmers in the benefited area who make drainage and other investments, and who otherwise intensify production. However, these several effects are essentially joint economic products or services of the total project investment. Unless a structural improvement is made for surface water control, improvements in sub-surface water control are of no use. If flooding and impaired drainage are joint problems in the watershed, SCS economists are cognizant that enhancement benefit 5Enhancement benefits incorporate effects that may be attributed to flood prevention or drainage or irrigation or any combination of these. SCS separates them from FWDRB for policy reasons, although this presents some conceptual problems, as indicated in chi 2. FWDRB are recomputed as a part of redefined enhancement benefits in ch. 7, in a study of 12 Mflchigan PL 566 projects. Project structural (SCK) and associated (ACK) investments are insufficient to achieve these benefits, for farm managers are assumed to intensify and change land use. That is, ordinary crop inputs (seeds, fertilizer, chemical sprays and other inputs) are assumed to be used at an increased rate; crepping patterns may be changed; and previously uncropped land may be crOpped. Of course, SCS deducts the costs of these changes, but their effect on the value of output still requires that they be completed, as assumed by SCS. 52 separations can't be made through deduction from observed values, especially in flatland areas.6 As a matter of agency policy a 50:50 division was used for the example project: Flood prevention benefits (FPB) = FWDRB + 1/2 MILUB + 1/2 LUCB; Drainage benefits = 1/2 MILUB + 1/2 LUCB.7 Even if FWDRB are not separated, such as for watersheds where channel work only is planned (i.e., no floodwater retarding structure, meaning a dam with a temporary, flood-holding reservoir), this policy-based separation may be made if flooding and impaired drainage are joint problems. The relevant computational routine and assumptions are specified in column 3, Table 3.1. While the necessary discussion is too extensive for this overview, it should be pointed out that, besides separating benefits, SCS procedures may tend to emphasize FWDRB, which are project-credited increments in net farm income only, and may tend to de-emphasize enhancement benefits, which are project-credited farm income, as re- duced for partial and delayed achievement of the with—project level of output. This will be considered in more detail in chapter 7. The policy context of the emphasis on FWDRB is given in chapter 2. Crop price, cost, yield and other assumptions are discussed in chapter 6. fi'vf 6Interview with John L. Okay, economist with the SCS Planning Party in Michigan, June 1969, on the topic of project evaluation in flatland areas, such as MHchigan. 7For the example project a benefit-based allocation of costs to flood prevention and drainage was used, but current (1970) SCS regu- lations require separate benefit and cost allocations (see ch. 2). 53 V’fi FWDRB and_§ydrology8 The SCS Economics Guide prescribes four methods of estimating the economic value of flood reduction: one of these has already been discussed as the net-income change analog of FWDRB, and another is similar, but for areas without defined stream channels. The two remaining methods are alternative ways of estimating FWDRB, the chief difference between them being the way in which the SCS hydrologist determines the expected extent of flooding (in physical terms). In both of these methods, a damage-frequency curve is developed: several convenient probabilities of occurrence are selected, related damage values are computed, and the paired probabilities and damage values are used as plotting points for the continuous curve, with the area under the curve representing expected annual damage. Plotting-point floodwater Composite acre value Plotting—point damage (FWD) = (CAV, typical acre x acreage flooded loss value in the (AF) floodplain) In relating acreage flooded to the selected plotting-point probability of occurrence, the SCS hydrologist uses either: (1) In the storm-rainfall frequency method, flood data are based on intense rainfall (storm) event frequencies of occurrence, related 8References for this section include: Harold O. Ogrosky, "Hydrology of Spillway Design: Small Structures?- Limited Data," Journal of the Hydraulics Division, ASCE (American Society of Civil Engineers), vo1.§0,‘nb. HYB, Proceedings Paper 3914, Mby 1964, pp. 2956310. Also, see Harold O. Ogrosky and Victor Mbckus, "Hydrologyof_Agricultural Lands," sec. 21 in Ven Te Chow, ed., Handbook.of Hydroldgy: A Compendium of Water Resource Technology (New YorkziMcGrawaHill Bock Co., 1964). Harold Ogrosky is Chief, Hydrology Branch, Engineering Division, SCS, USDA, washington, D. C. Victor Mockus is the author of the SCS HydrolOgy Handbook, NEH-4 (see note 4). Also, see Ven Te Chow, "Statisticalvand Probability Analysis of Hydrological Data," sec. 8 in Ven Te Chow, ed., Ibid. 54 rainfall data, watershed measurements, and some assumed storm and watershed conditions. This method is widely used by SCS, owing to the lack of historical flood data for.most small watersheds. (2) In the historical method, flood data are based on actual time series data for such variables as peak floodwater discharge rate (measured in cubic feet per second), related water-surface stage (water-surface elevation in feet, as measured at a stream gaging station), and related pgint rainfall (measured at a nearby recording rain gage for a geographic point, hence the name point rainfall data). Newspaper accounts, actual measurements, and local residents may be called upon for information on the extent of flooding, and this is related to the frequency data for the historical rainfall or stream gage data. Statistical and probability concepts are used in both methods. The terms probability of occurrence, frequency of occurrence and return period all refer to the same concept, and relate to continuous variables (not discrete variables), and continuous statistical frequency distribution functions (not discontinuous or step functions). For example, for the location of the North Branch of Mill Creek water— shed, Mflchigan, the 25-year return period, 6-hour intense rainfall is 2.90 inches (point estimate), according to the Weather Bureau reference map. In other words, there is a 4% chance (ex ante) during any one year that, for this location, the 6-hour duration intense rainfall will squalor;gxceed 2.90 inches, where the annual, ex ante probability of occurrence, P = l / (return period in years) = 1/25 = 0.04 or 4%. The same concepts may be applied to damage, stage, acreagenflooded, rainfall and other variables. In the storm-rainfall frequency method, 55 intense-rainfall event frequencies are assigned to all of these other variables, given the relationships and assumptions of the SCS model. The selected return periods are the l, 2, 5, 10, 25, 50 and sometimes 100 year return periods, and the associated annual, ex ante probabilities of occurrence are 1.00, 0.50, 0.20, 0.10, 0.04, 0.02 and 0.01, respectively. For engineering design purposes, rainfall, stage and discharge amounts of unstated, but implicitly much smaller probability of occurrence are used. Suffice it to say that flood damage estimation and engineering design criteria development are concerned with opposite ends of the flood or rainfall frequency distribution, roughly speaking.9 SCS (Storm-Rainfall) Frequency Method This method employs Weather Bureau intense rainfall event, that is storm rainfall event data, and some rather complex relationships and assumptions to develop peak floodwater discharge rates. More discussion, a critique and sensitivity analysis are presented in chapter 5. Briefly, the process is as follows: 9Special Weather Bureau studies have been commissioned by SCS and the Corps of Engineers for the purpose of studying what is called probable maximum precipitation (PMP) for all parts of the United States. For the location of the North Branch of Mill Creek watershed, the 100- year, 6-hour, point rainfall estimate is 3.50 inches; the 1—year, about 1.50 inches; by interpolation, the author estimated the lOOOéyear (P = 0.001) amount as about 5.2 inches; but the probable maximum precipitation is 24.0 (twenty four) inches! Design floods for various components of a dam may be based on 25-year, 50-year, lOO-year, or some combination of lQO-year and probable maximum precipitation (which has no explicit probability assigned, except to say that it is extremely rare). Corps of Engineers designs may be based on what is called a standard project flood which is based on lowwprobability rainfall, assumed watershed conditions and some observations of actual floods in the region. 56 Step 1. For the specific watershed location, storm duration, and selected probabilities of occurrence, determine the intense rainfall amounts from the Weather Bureau reference, TP:49310 Adjust these point rainfall amounts downward, if the watershed area exceeds 10 square miles, the diminution being proportional to watershed area. The storm duration, in hours or minutes, is equated approximately with the watershed time of concentration (To), which is the time required for water to travel from the most distant point along its natural course to the watershed outlet. Step 2. Using the SCS rainfall-runoff relationship, determine the depth of runoff (in inches) for the several storm rainfall depths. In this step, watershed soils are classified into one of four hydrological soil groups; plant cover types are surveyed; ground lepe and other conditions are determined; and, assuming "average" soil moisture levels (AMC-II), and mid-growing season plant growth, the appropriate runoff curve number (CN) is selected. The higher the runoff curve number, the greater the depth of runoff for any given amount of rainfall. In terms of variables, barren land will produce more runoff than heavily pastured or wooded land; coarse (sandy) soil permits more infiltration and See U. S. Weather Bureau, Rainfall Frequency Atlas of the United States for Duration from 30 Minutes to 24 hours and Return Periods ‘ fromfl to lOOTYears:‘TechnicaljPaperfiNo. 40, by.David.M. Hershfield (WashingtOn, D. C.: USGPO, ;May 1961); commonly called TP—40 by SCS. Or see U. S.-Weather Bureau, Two to Ten Day Precipitation for Return ‘ Periods of 2 to 100 Years in the Contiguous United States, Technical 'Paper No. 49, by John F LMiller (Washington, .D. C.: USGPO, 1964); commonly called TP-49 by SCS. These and other studies were specially commissioned by SCS 57 therefore less runoff than heavy (fine-particle, clay) soil; and wet or frozen ground lets rainfall runoff, while dry ground allows infiltration. Given the adjusted storm rainfall amounts, go to the proper table or graph for the determined runoff curve number, enter at an indicated rainfall amount and read off the related runoff amount on the other axis. Repeat this process for each of the rainfall amounts. Step 3. Rate the stream channel and valley at various points along the stream length. For this step, valley and streamrchannel cross-sectional profiles, slopes (stream gradients), channel roughness, and other factors are determined by field measurement and inspection. Various rates of flow are compared with the resulting cross sectional channel-valley capacity, and a stage-discharge curve is drawn, showing the water surface stage (elevation in feet) for various discharge rates (in cubic feet per second). Secondly, the stage-area inundated relationship is developed for each hydrological reach. Step 4. Taking into account the runoff depths in the watershed for the selected series of storms, and the drainage area contributing to each hydrological reach, route each of the resulting floods progressively downstream through all reaches. The water volume (in cubic feet) is represented by a triangular hydrograph's area, with time measured along the triangle's base, and rate of flow measured perpendicularly (vertically) upward from the base (in cubic feet per second). Each hydrological reach has an inflow hydrograph from the reach immediately upstream, and an inflow hydrograpb.for any sidestreams and local runoff occurring within the reach. The flows are added. The peak rate of 58 flow, that is the rate for the peak of the hydrograph, is of interest according to the unit hydrograph theory, which was developed some 30 years ago.11 The peak floodwater discharge rate for each of the selected storms is entered on the stream rating curves for the hydrological reach in question, and the stage is read off. In turn the stage-area inundation curve serves to convert stage to area inundated. Summagy: For estimating FWDRB and flood damages on an expected annual basis, this process yields a series of paired acreages flooded and frequencies of occurrence. While the frequencies are basically for rainfall amounts produced by storms, the relationships and assumptions of the SCS model make them applicable to the acreages flooded and damage amounts.12 The series of paired frequencies and damage amounts are used to plot the SCS damage-frequency curve for each economic reach and for both the with and without project situations. The differences in damage constitute the FWDRB. SCS Historical Method (for Time Series Flood Data) The historical and frequency methods are similar in many respects, as already indicated. Both involve the development of damage-frequency llFor discussions of the methodology, see for example; Chester 0. Wisler and Ernest F. Brater, H drolo , second edition (New York: John Wiley and Sons, Inc., 1959); Daniel W. Mead, Hydrology--The ' Fundamental Basis of Hydraulic Engineering, second editionTCNew York: McGraw Hill Book Co., 1950). 12These assumptions are studied in Ch. 5. 59 curves. The basic difference is that in the historical method the paired series of damage values and frequencies are based on hydro- logical analysis of a time series of historical peak floodwater discharge data for the stream in question, rather than on simulated or synthesized discharge rates and stormaevent (intensesrainfall—event) frequencies of occurrence. Since stream gaging stations are rarely located at ideal Spots for a proposed project, data must be developed. The transient nature of economic values, that is their non-homogeneity through a time series of damage data, makes it necessary to develop data for acreages inundated only. As in the frequency method, the damage done by a particular extent of flooding is therefore determined as the product of acreage flooded and the composite acre value (typical acre loss value) in agricultural areas. One difficulty with this method is that the time series of discharge data may not represent homogeneous hydrological conditions, such as if man-made developments have altered the stream, or if the valley plant cover has been changed, such as from forests or other native growth to cultivated crops, or from pasture to row crops, or from agricultural to urban uses. SCS Modified Historical Method In practice SCS uses combinations or variants of the storm— rainfall frequency and historical (time series flood data) methods. A simplified historical method is used to estimate pppfcrop enter- prise damages for small watersheds; this relatively unimportant item consists of damage to roads, bridges, railroads, utility lines, buildings, farmsteads and other property in rural areas. Local 60 people, old newspaper accounts, and recorded flood series for nearby streams provide a basis for estimating two or three plotting points for a rough damage-frequency curve. FWDRB Estimation: Computational Steps The estimation of agricultural FWDRB (floodwater damage reduction benefits) involves several computational steps that take into account flood, watershed and crop-enterprise variables and assumptions, both hydrological and economic. The computational process leads to a set of flood damage values, each of which is paired with an ex ante probability of flood occurrence. These paired values identify plotting points for the SCS damage-frequency curve (Figure 3.1). In agricultural floodplains, the chief source of flood loss is associated with crop enterprises which rank far below urban residential, commercial and industrial property in terms of potential loss. However, the damage estimation process in an ex ante sense is more complex for crop enterprises, because both the values subject to loss and the probability of flood loss vary during the growing season. Monthly, 100% Flood Loss Values: Step 1 In this step, two-week loss values are averaged into monthly loss values (accounting for the factor 0.5 in the following algebraic formulation). The initial assumption of 100% flood loss is a computa- tional convenience, and means that the growing crop is completely destroyed (late season losses), or that whatever has been done in the way of cropping practices must be repeated (early season losses). The 100% flood loss (FDM) has the following more specific meaning. Early in the season, replanting costs (RPC) represent the only loss. 61 Later, a yield reduction occurs, because of late planting, resulting in an added loss. Still later, the gross value of the original crop in the field (P - AVC, price less average variable cost), less avoided costs (AVDC), represents the loss. If a substitute crop can still be planted, the original crop loss is reduced by whatever can be gained in the form of net returns. SCS assumes that about two weeks are required to allow fields to dry sufficiently to permit normal cropping practices to be performed. The following algebraic formulation is designed for a computer, and all variables have a value specified for each two week period. SCS economists usually perform the computations for without-project conditions only (j = 1, see note 15). 2 FDMm,k,j = i i0,5 x PUNHi,m,k x [PCi,m,k x Yk,j x (Pk - AVC k) ' AVDCi.m,k ' PCAi.m.k x Yks,j x (Pks ' AVCks) + RPCi.m,k.j]} Briefly the variables are defined as follows:13 FDM: monthly, 100% flood loss value, per—acre, by crop and crOp production intensity level. PUNH: portion of the crop unharvested PC: portion of the original crop expected not to yield; see PCA; the complement of PCA for any given crop. Y: crop yield per acre. P: crop price per unit of output. AVG: crop variable cost per unit of output. fir l3Variable names and subscripts are used consistently in this description of FWDRB and enhancement benefits. Illustrative computations for FDM for the 16 twosweek.periods in the growing season are shown in Vondruska, pp. 11-12. 62 AVDC: noanVC avoided cost per.acre. PCA: portion of the substitute crop expected to yield; see PC. RPC: replacing cost per acre. subscripts are: m: month; m = 1, ..., 8. k: crOp; for this project k = l, ..., l7. ks: substitute crop; numerically, ks # k. i: one of two two—week periods in a month. j: watershed condition, either without the project (j = 1), meaning flood—free but poorly—drained conditions, or with the project (j = 2), meaning flood-free and well-drained conditions; the associated production intensity levels. Example project computations of FWDRB assume j a l. By-Crop Annual Loss Values: Stgp 2 In this step the 100% monthly loss values obtained in step 1 as an average of tw0dweek loss values (the averaging process is summarized in Table 3.2) are weighted and adjusted to obtain the annual loss values. Two operations are involved. Given the 100% monthly flood loss values, the FDM's, it is necessary to adjust for the effect of limited destruction in terms of depths of inundation, of which there are two for the example project. To form an annual per—acre loss value (CFD) for a given depth of inundation (id), crop (k) and level of crop production intensity (j), weight the monthly values (FDM's) by the monthly probability of flood loss occurrence (PM) and the depth.adjustment factor CD): CFDid,k_,j = A (FDMm,k,j X Dm,id,k X PMm)~ 63 Table 3.2.. Loss Value Computations, Corn for Grain. _‘. __‘, Item April May June July Aug Sept Oct Nov Annual Loss 1 $4.45 7.30 31.70 70.50 70.50 70.50 52.88 3.53 Loss 2 $4.45 22.38 69.16 70.50 70.50 64.86 14.10 «0n Value, or Average monthly, 100% loss, FDM value, (loss 1 + loss 2)/2, $'s. CFD, 4.45 14.84 50.43 70.50 70.50 67.68 33.49 1.76 for corn, Depth adjustment factor, D for 0-2 ft., D for 2+ ft. inundation in 1 -0- .50 .75 .64 .42 .22 .28 .34 $'s. 2 -0- .55 1.00 1.00 .80 .55 .60 .70 Monthly probability of flood loss occurrence, PM. .05 .26 .21 .16 .ll .05 .16 -0- Weighted monthly losses; annual losses (CFD's), two depths, $'s. l -0— 1.93 7.94 7.22 3.26 .74 1.50 -0— 22.59 2 -0- 2.12 10.59 11.28 6.20 1.86 3.21 -0- 35.26 Source of original data (modified slightly): USDA, SCS, documentation for the North Branch of Mill Creek watershed, Michigan. Output and input prices are on a projected long term basis, using 1960 data. This data is for corn for grain (corn for silage is treated as a separate crop) on a per-acre basis. Composite Acre Values: Step 3 The annual loss values, CFD's, for each crop from step 2 are multiplied by the proportion of the floodplain planted to that crop (R), and the arithmetic products for all crops are summed to form the composite acre loss value (CAV): _ Z CAvid,ir,j — k (CFDid,k,j x Rk,is,ir)' The CAV's are estimated for depths of inundation (subscript id), economic reaches (ir), and levels of crOp production intensity (j). Selected planting pattern data (R) are shown in Table 3.3. In II II H H 0 computing FWDRB the situation subscript ( is in Rk,is,ir) is specified in the computer subroutine to obtain the proper cropping pattern, as discussed elsewhere (see Vondruska, pp. 57-58). For 64 the example project, SCS used only the CAV's for the lower level of crop production intensity, necessitating an adjustment in the enhancement benefits (see FWDC in Table 3.7).14 Table 3.3 Composite Acre Values, Depth 1. Crop Proportion of the flood Annual loss value Summation zone in this crop, for this crop, Rk,is,ir CFDid,k,j Corn 0.082 x $22.59 = $1.85 Wheat 0.065 x $17.17 = $1.12 Potatoes 0.238 x $121.16 = $28.83 Other crops . . . . . $69.72 Total, or composite acre value, CAV Source: USDA, SCS, documentation for the North Branch of Mill Creek watershed, Michigan, February 1962. Output and input prices are on a projected long term basis, using 1960 data. For economic reach 1 and depth 1 (0-2 ft. inundation). Estimated Flood Damages: Step 4 Expected annual floodwater damages (FWD) are computed for both with and without project conditions in the watershed (subscript it), by economic reach (it), given the composite acre values (CAV's, typical acre loss values) for the floodplain area, and the plotting-point acreage flooded (AF). Damage values are computed for each of the selected probabilities of flood occurrence; the related pairs are used 14The computational routine used by SCS to compute FWDC is given in Vondruska, p. 56, and in note 4, ch. 6. 65 as plotting points for the damage-frequency curve (Figure 3.1). Alternatively, the area under the curve may be approximated as the sum of rectangular areas (Figure 3.1), with one area for each plotting point (ist = l, ..., 6 for this project): _ Z Z FWDit,ir - istFwist Iid (AFid,ist,it,ir x CAVid,ir,j)] Value of damage done by one extent of flooding. Damage represents the height of the rectangle, and selected-flood weights (FW's, see Table 3.4), the width. Adjusted Flood Damages: Step 5 For the example project an upward adjustment was made in the floodwater damage (FWD) value obtained in step 4 to take account of flood recurrence during the growing season and the difference between the largest or most extensive flood and the most damaging flood, as shown in Table 3.4. These adjustments are significant, but an explanation requires more background in hydrology than can be presented here; further discussion is deferred to chapter 5, except to note that there may be implicit as well as such explicit adjustments in the underlying hydrological data. FWDRB: Step 6 The withrproject and without—project flood damages are compared, and the difference represents floodwater damage reduction benefits Withoutvproject flood damages, reach 1 $60,055 Withrproject flood damages, reach 1 5,177 FWDRB $54,878 Figure 3. 140 120 100 (O .0 O O H W 80 d -H Q) 00 CO E (U 'U 3 60 3.) CO 3 'U 0 O H Lu 40 20 0 Source: 1. SCS Damage Frequency Curve. 66 l x Expected annual probability of occurrence Table 3.4. 0.4 l 0.6 0.8 1.0 67 Table 3.4 Annual Probable Flood Damages. Rectan- Weighted Acres Per-Acre gular contri- flooded damage Unweighted area bution to Flood Annual two or loss, damage, approx. expected return proba- depthsb annual dollars weight annual perioda bilitya (AF) (CAV) (AF x CAV) (FW) damage 50-year 0.02 1863 $69.72 $129888 155 74.32 11520 $141408 0.02 $2828 25-year 0.04 1702 $69.72 $118663 97 74.32 7209 $125872 0.05 $6294 lO-year 0.10 1430 $69.72 $99700 25 74.32 1858 $101558 0.08 $8125 5-year 0.20 1080 $69.72 $75995 -0— 74.32 -0- $75995 0.20 $15199 2-year 0.50 443 $69.72 $30886 -0- 74.32 -0- $30886 0.40 $12354 l-year 1.00 35 $69.72 $2440 -0- 74.32 —0- $2440 0.25 _§§lg- Total . . . . .1.00 $45410 Adjustment for flood recurrence (1.15) and for the "most damaging versus the largest" floods (1.15), or (1.15 x 1.15 =) 1.3225 Probable annual damage, without-project, adjusted . $60055 Source: USDA, SCS, documentation for the North Branch of Mill Creek watershed work plan. term basis, using 1960 data. project. Crop input and output prices on a projected long Data for economic reach 1, without the aExpected annual probability of occurrence (P) equals 1 / (return Return period and recurrence interval (years) are synonymous. All refer to points along the SCS damage frequency curve period, years). (as in Figure 3.1) for the continuous variable, damage. P = 0.04 that damage will equal or exceed $125,872. For example, Probability (P) is sometimes called an exceedance probability by hydrologists. bDepth 1, 0-2 feet inundation; depth 2, inundation over 2 feet. 68 Given the computation—reducing, simplifying assumptions used by SCS, the model does not take account of_managerial—reaction to the reduction in flood hazard on the estimated FWDRB, $54,878. The two estimates of flood damage are for projected, but without—project economic conditions in the watershed, with the difference in damages, that is the FWDRB, being due to the difference in expected annual physical extent of flooding. In other words, the same composite acre values (CAV's, or typical acre loss values) for the floodplain have been used throughout the computational process for the example project.15 FWDRB Estimation: Summary FWDRB (floodwater damage reduction benefits) account for only a part of total project benefits. According to the assumptions used by SCS for the example project, FWDRB do not reflect the economic activity (management) effect of reduced flood hazards, for FWDRB essentially 15Managerial, cropping pattern and other changes for the with-project economic condition of the watershed would result in a higher CAV. Obtaining a second CAV set is quite a computational burden; therefore, SCS approximates the effect in aother way (see note 4, ch. 6 and Vondruska, p. 56, on FWDC computations). With—project damages (FWDz) are higher than the $5,177 indicated here, by $1,223). Thus, FWDRB are $1,223 too high“ SCS takes this difference into account, not by reducing FWDRB, but by reducing enchancement benefits (see the FWDC deduction, $1,223, in Table 3.7). However, the two deductions are not equivalent in effect on total project benefits; a deduction of $1,223 from FWDRB would have a greater impact, because the deduction from enhancement-benefits net income is actually reduced in subsequent manipulations (in obtaining the data in col. 7 from that in col. 6, Table 3.7). Given the assumptions about achievement of the withnproject level of farm-income, as used by SCS, this procedure seems acceptable.' However, the SCS assumptions are discussed further in ch- 7. 69 represent the effect of projectvcaused differences in the expected annual extent of flooding in purely physical terms on the projected (futuristic), but without—project economic condition of the watershed. Thus, in this case, the same economic values, as reflected in the composite acre values (CAV's or typical acre loss values) for the floodplain, were used by SCS to compute both with and without-project floodwater damages (FWD). Hence, the difference in damages, FWDRB, is due solely to a reduction in the physical extent of flooding. However, reductions in flooding result in a variety of changes in watershed economic activity, and in increases in net income in addition to those counted as FWDRB. Thus, reduced flood hazards are assumed to cause farm managers to shift to a higher crop production level, and, in the case of Michigan projects, to make additional capital investments in land improvements, such as for drainage, land clearing, field leveling and shifts from non-crop to crop uses of farm land. The net income changes forthcoming are called enhancement benefits, which are in addition to FWDRB, and they will be considered in the next section. The separation of the two is made for policy reasons, and depends upon hydrological data on flooding in the case. of FWDRB, although a FWDRB-equivalent could be estimated as a net income change, much like enhancement benefits in terms of computational routine. Enhancement Benefit Estimation: Computational Steps As shown in Table 3.1 at the beginning of this chapter, overall project benefits may be obtained as the difference between net farm in— come with the project and without the project for the affected part of the 70 watershed, but, for policy reasons, SCS separates various categories of benefits. FWDRB (floodwater damage reduction benefits), as discussed in the preceding section, constitute one form of net income improvement, although they are conceived conversely as damage or loss reductions. Of course, notwithstanding the consequent policy implications, project benefits as a whole could be conceived and computed either as damage reductions (due to removal of excess water problems, meaning poor drainage or flood problems, or to the removal of other limitations to intensification crop production or cost reduction), or, alter- natively, as enhancements (due to production intensification or cost reduction). In any event, in the computational and conceptual alternative chosen by SCS for the North Branch of Mill Creek watershed, Michigan, and for most projects in Michigan, FWDRB are computed as the difference in floodwater damages associated with the reduction in the expected annual physical extent (acreage) of river overflow on the watershed in its without-project, but projected (futuristic) economic condition. A net-return-change equivalent of FWDRB may be computed: Net income change Net income under Net income with due to flood = flood-free & poorly- - flooding & poor damage reduction drained conditions drainage alone To FWDRB or the net return equivalent of FWDRB.must be added the net return improvements, which.SCS calls enhancement benefits (EB) and which are associated with_more intensive land use CMILU) of already cropped farm.land, and with.land use change (LDC) for previously uncropped famm land. This second class of income improvements is, conceptually, the result of management reaction to reduced flood hazards, 71 improved drainage outlets, and new irrigation water supplies, all project effects: Net income change Net income under Net income under due to more inten- = flood-free & well- - flood—free & poorly— sive & changed drained conditions drained conditions land use An example of the computational details for this second kind of benefits, enhancement benefits, follows. Computing the Net Return Change: Step 1 Flood hazard reduction makes possible changes in farm practices and investments, both of which can increase net farm income. One subclass of these enhancement benefits (BB) is called more intensive land use benefits (MILUB). For example, flood relief can allow the farmer to switch from lower valued to higher valued crops, apply variable inputs (seed, fertilizer, sprays, etc.) at a higher rate, and perform crOpping practices in a more timely fashion. (Planting, spraying, harvesting and other activities must be delayed for days-— SCS assumes l4 days--after a flood or heavy rain.). With the flood hazard reduced, drainage outlets provided, and irrigation water supplies provided, all largely at Federal expense via a PL 566 project, managers may find it profitable to drain or irrigate their land, typically taking advantage of ACP payments, another Federal cost (that would not likely be incurred without the PL 566 project). A second class of enhancement benefits, land use change benefits (LUCB), are computed in a similar fashion, and they relate to land that had been uncropped without the project. The project mainstream works make it economical to clear some land formerly in woods, or to 72 prepare and plow land that was formerly in permanent pasture and idle uses. The value of the former production of LUC land is usually counted as zero. Net returns are computed using the following formulation. Illustrative computations are shown in Table 3.5, although the order of arithmetic operations is changed. NRir’iS = i {[vk,j x (Pk - AVCk) - AFCk,j] x ABir,iS x Rk,is,ir} / / / Net returns Per-acre net returns for one The acreage planted (for is = 1, cr0p, using the yield and to one crop jsl; for costs sets for either the is = 2 or 3, lower (j = l) or higher j 8 2) (j = 2) intensity of crop production. The variables are (for details, see Vondruska, pp. 54—58): NR: net returns Y: crop yield per acre. P: crop price per unit of output. AVC: average variable cost per unit of output. AFC: non-AVC costs, per acre. AB: acreage benefited (see Table 3.6). R: ‘portion of an area planted to a crop (see Table 3.6). Subscripts are: it: economic reach. is: situation (see Table 3.5). k: crop. j: crop production intensity level, either without the project (j = 1), under flood-free and poorly-drained conditions, or with the project (j = 2), under flood-free and well-drained conditions; refers to yields and related costs. 73 “Noma >uwsunmm vmumv .smwanofiz .vonmumums xmmuo Hafiz mo Seaman nuuoz How soaumusmssoov .mom .< H M m umoo saws» use: umoo moans aouo .H gamma .Amzv wauaumm umz wasasmupo .m.m magma 74 As shown in Table 3.5, without-project net returns are $294,335. They are the same as MILUB net returns NR1 and are computed assuming flood-free, but poorly-drained conditions (using Yland AFCl data). MILUB net returns NR2 are $451,435; they are computed assuming flood- free and well-drained conditions (using Y2 and AFC2 data). ‘MILUB net returns are computed for the previously cropped area (3663 acres). LUCB net returns are computed for the previously uncropped acreage (332 acres); they are $38,046 (based on Y2 and AFC2 data). This assumes a zero value for the output of this land in the without-project condition. The area used (332 acres) was obtained by SCS, as shown in Table 3.6. Not all of the net return increases developed in this step are assumed to occur; downward adjustments are discussed in step 3. Table 3.6. Obtaining LUCB Acreage, Economic Reach l.a Land use Without-project Conversion LUCB With—project acreage for LUCB acreage acreage Crops 3663 --------------- 3995 (3663 + 332) Permanent pasture 50 50 x .90 = 45 5 Idle farmland 40 40 x .90 = 36 81 4 Farm woods 335 335 x .75 = 251 84 Miscellaneous 160 ------ - - 169_ 4248 332 4248 Source: USDA, SCS, documentation for North Branch of Mill Creek watershed, Michigan, dated February 1962. 8In computing NR1r is for economic reach 1 (ir = 1), ABi is = 4248 acres for is = 1 add is s 2 (for MILUB) and ABir is = 33 'acres for is s 3 (for.LUCB). For more details, see VondruSka, pp. 54—58. 75 Deductions:f Step 2 The net returns computed in Step 1 are summarized in Table 3.7, and the change (col. 3) is computed. As previously indicated, enhancement benefits (EB) are partially net benefits, because associated costs are deducted (including both ACK and ACOM components, Table 3.7, col. 4). Recall that no such deduction was made from FWDRB (floodwater damage reduction benefits). Also note that structural costs (both SCK, capital, and SCOM, operation—maintenance components) are counted in the denominator of the SCS benefit cost ratio for evaluating PL 566 projects; these costs will be considered shortly. The second deduction, FWDC (Table 3.7, col. 5) relates to the assumptions used by SCS in computing FWDRB. For this project FWDRB Table 3.7. Summary of Net Returns and Enhancement Benefits. Enhancement NR Assoc. benefits Item NR1 NR2 change costs FWDC EB-lOOZ (EB)a $ $ $ $ $ $ ‘71 2 ’3 4 5’ 6 *7fi Compufed as: W'j'ffi :VZ-l 344¥5 For MILUB on previously cropped 3663 acres 294335 451435 157100 35904 985 120211 85707 For LUCB on previously uncropped 332 acres Total '0? ‘ 38046 38046 5701 238 - - - - — - P,I -0- 9283 9283 865 58 8360 7106 Woods -Oe 28763 28763 4836 180 23747 17431 Total 294335 489481 195146 41605 1223 152318‘ 110244' Source:T'USDA:fSCS,‘documentationffor North Bfanch.of.Milljcreek watershed, Michigan, dated February 1962; reach 1. aEnhancement benefits in column 7 are obtained from the EB-lOOZ data in column 6 as shown in Table 3.8. 76 were computed using a single composite acre value (CAV, based on Y1 and AFCl data), with these benefits due solely to the project- credited decrease in acreage flooded. Farming at the higher, with— project output levels (based on Y2 and AFC2 data) increases with- project damages (FWDz) above the level assumed in computing FWDRB.16 As previously indicated in this chapter, and as will be considered in the sensitivity analysis of chapter 7, these deductions may be interpreted as serving to emphasize FWDRB and to de—emphasize enhancement benefits (EB). Data for LUCB in Table 3.7 reflect the SCS division (see Table 3.6) into portions for previously uncropped land formerly in pasture and idle uses, and in farm woods. Obtaininngnhancement Benefits (EB): Step 3 After the enhancement benefit (EB) component of project-credited farm income has been reduced for associated costs and possibly other items, there remains what the author calls EB-100% ¢he data in Table 3.7, col. 6). However, SCS typically assumes that the with-project level of farm income (NR2) correSponding to this EB-100% will be only partially achieved over a period of 15-20 years. In terms of cash flow concepts, a certain portion of the EB—lOOZ annual rate is assumed to accrue in each year in the evaluation period. Typically, SCS assumes that these annual EB cash flows increase in the 16What the author calls FWDC (floodwater damage change) SCS calls "adjustments for remaining flood damage to higher values"; see Economics Guide, ch. 4, pp. 13«l4; SCS Table 4.4 there is similar to a combina— tion of the author's Tables 3.5 and 3.7, and includes a deduction for this item. 77 approximate fashion of a decreasing-rate growth curve (concave from below) over a 15v20 year period, after which they remain constant, usually at less than the EB-100% rate. Actually, SCS uses a linear (rectangular), segmented approximation of a growth curve, with three, straight-line segments for the example project, as shown in Figure 7.3. For the example project, counting from time zero, when construction is initiated, 35% of the EB-100% rate is achieved by year 5, after a S-year installation period; an additional 25% by year 10;17 and another 20% by year 20, for a total of 80% for years 20-50. However, SCS typically assumes "instant installation";18 thus 35% of the EB-100% rate is achieved at time zero (rather than by year 5), 60% by year 5 (instead of year 10), and 80% by year 15 (rather than by year 20). The differences between EB growth polygons are shown in Figure 7.4. As described in chapter 7, cash flows are computed by the author for each year. Yet, SCS economists typically do not compute cash flows for each year. Visualizing the series of cash flows as a polygon, as in Figure 7.3, the SOS short-cut involves slicing it into a few horizontal segments, rather than into vertical segments (one vertical segment for each year). A horizontal segment's portion of EB-100% is first computed, then a "discount factor" is applied, along with a fit f ‘- v w 17Data for Mill Creek for MILUB.only, see Appendix, Table 2. 18See WPH, sec. 102.0211; Economics guide, ch. 4, p. 13, is more conserVatiye. The effects of discarding‘this and other related assumptions are studied in chapter 7 for 12 Michigan PL 566 projects. 78 further adjustment for distant—time segments. An illustration is shown in Table 3.8.19 Table 3.8. Computation of EB from EB—100%, MILUB Data. Item ' Time zero ' Yrs. 1-5 ‘ Yrs. 6e15 EB—100% $120,211 $120,211 $120,211 Segment % X.35 X.25 ‘ X.20 $ 42,074 $ 30,053 $ 24,042 Discount (None) For 5 yrs. For 10 yrs. factor - - — X.914 X.818 $ 42,074 $ 27,468 $ 19,666 Present value of 1, 5 yrs. hence X.82193 $ 16,165 Sum ($42,074 + $27,468 + $16,164 =) $ 85,707 Source: USDA, SCS, documentation for the North Branch of Mill Creek watershed, Michigan, dated February 1962; MILUB data only, for reach 1; interest rate of 4%, evaluation period of 50 years. The EB amount for MILUB at the bottom of Table 3.8 is the same as that shown in Table 3.7 (col. 7) and Table 3.9. Enchancement Benefit Estimation: Summary In the preceding section, floodwater damage reduction benefits were computed as the effect of reduction in the extent of expected annual acreage flooded in physical terms for the without—project, projected (futuristic) economic condition of the watershed. Enhance“ ment benefits, covered in the present section, represent the additional 19The discount factor for years 6-15 should be .6431, as computed in accord With Economics Guide, Appendix A, rather than .6723 (.818 x .82193), as shown in Table 3.8. This error affects the author's computations in chapter 6, and many SCS computed benefit cost ratios in Appendix, Table 3. Discussion with John Okay, economist with the SCS Planning Party in Michigan, in October 1970, indicated that this error has been recognized and that it no longer affects SCS evaluations. 79 income made possible by the complementary: (1) main project works; (2) on—farm and inter~farm (associated cost) works; and (3) management choices which intensify crOp production. Project effects on farm income will be reviewed later in this chapter. The process of totalling all of these benefits, allocating them to officially designated purposes, and obtaining the benefit cost ratio will be considered following a brief discussion of project costs. Project Costs Project costs have already been discussed in chapter 2, and the presentation here is related to their use in formulating the benefit cost ratio. Structural costs for the mainstream project works are counted in the denominator of the ratio. Associated costs for the complementary on-farm and inter-farm works are deducted only from the enhancement-benefit, project—credited farm income, not from the project-credited farm income counted as FWDRB. Both structural and associated costs have capital and operation—maintenance components; both include Federal and local items of cost, if ACP (Agricultural Conservation Program) payments are counted as an item of Federal cost. Several interest rates are used in the computations.21 21Costs for the example project are shown in Table 2.2; those for 12 Michigan PL 566 projects in the Appendix, Table 2, where the example project is cited as the Mill Creek. Abstracting from Table 2, in the Appendix and from discussion in ch. 7, for the example project, costs were amortized as follows: structural capital costs (SCK), at 2 5/8%, over 50 years; associated capital costs (ACK), onsfarm, at 6%, over 50 years; and associated costs (ACK), internfarm, at 5%, over 50 years. A fourth rate, 4% was used to compute enhancement benefits. The single—rate equivalent of these multiple rates used by SCS is about 3.0%; that is, this rate will produce about the same benefit cost ratio (see Table 7.3, col. 4). 80 Obtaining the Benefit Cost Ratio Annual benefit and cost data are summarized and categorized, and the benefit cost ratio is obtained, as shown in Table 3.9. The benefit categorization into flood prevention (FPB) and agricultural water management (AWMB) is made for policy reasons.22 It may be useful to discuss the different kinds of FWDRB shown in Table 3.9. Crop and pasture FWDRB are the dominant kind of FWDRB for PL 566 projects in agricultural areas. To obtain total direct FWDRB ($79,787), other FWDRB must be added, and this includes the effect of reduced damages to farm and non-farm buildings, roads, fences, equipment, utility lines, and other property in the flooded area. Indirect FWDRB ($7,979) relate to inconvenience reduction in the area; they are not an estimate of property-damage-reduction benefits, but rather take into account the effect of avoided interruptions in business and service activity.23 2 Benefits were used by SCS to allocate costs for the example project, but other methods are now used. See chapter 2. 23808 computed indirect FWDRB as 10% of direct agricultural FWDRB; source of percentage, Economics Guide, ch. 3, pp. 31- 32. Suggested ranges are as follows: agricultural, 5- -10%; residential, 10-157; commercial and industrial, ISrZQZ; highways, bridges and railroads, 15—25%; and utilities, 15-20%. For agricultural direct damages, the "percentage probably will be much higher when irrigation and drainage facilities are damaged. The indirect damage should be determined on a case basis when these facilities are involved." 81 Table 3.9. Summary of Annual Benefits and Costs. fir ‘ _ f . Econ. Econ.. Reaches Econ.. .Project Kind of benefit reach I reach 2 1 and 2 reach 3 total Enchancement benefits (EB, from Table 3.7) . MILUB $85,707 $47,721 $133,428 $1,056 LUCB 24,537 40,485 65,022 " 1,278 Total $110,244 $88,206 $198,450 $3,314 Floodwater damage reduction benefits (FWDRB, from Table 3.4) FWD1 $60,055 $25,223 $85,288 ($106)a FWDzb 5,177 1,632 6,809 ( 2) FWDRB $54,878 $23,601 $78,479 ($104) Flood prevention benefits (FPB) Crop & pasture FWDRB $78,479 $78,479 Other FWDRB (property) 1,310 1,310 Subtotal (direct FWDRB) $79,787 $79,787 Indirect FWDRB (10%) 7 979 7,979 Subtotal (all FWDRB) $87,768 $87,768 FWDRB attributed to land treatment (estimated as 4% of above subtotal)b 3,512 3,512 $84,256 $84,256 MILUB FPB (1/2 of MILUB above) $66,714 $518 $67,232 LUCB FPB (1/2 of LUCB above) 32,511 1,139 33,650 Total FPB $183,481 $1,657 $185,138 Agricultural water management benefits (AWMB, not itemized in work plan) 1/2 of MILUB above $66,714 $518 $67,232 1/2 of LUCB above 32,511 1,139 33,650 Total AWMB $99,225 $1,657 $100,882 Total annual benefits (FPB + AWMB) $282,706 $3,314 $286,020 Annual structural costs (from Table 2.2) 37,372 3,148 40,520 Benefit cost ratio 7.56/1 1.05/1 7.06/1 Source; ’USDA, SCS, documentation for North Branch of Mill Creek watershed, Muchigan, dated February 1962. aBecause economic reach 3 crOp and pasture FWDRB were only $104, the SCS economist added them to MILUB.' bPositive valued item is deducted. 82 All FWDRB are totaled ($87,768), and a deduction is.made for the estimated effects of the flood-reducing land treatment practices. The basis of this deduction (the 4% shown in Table 3.9) is developed by the SCS Planning Party hydrologist.24 The remaining FWDRB ($84,256) are added to MILUB and LUCB to form the numerator of the benefit cost ratio. Before this final summation takes place, MILUB and LUCB are divided into FPB (flood prevention benefit) and AWMB (agricultural water management benefit) categories for policy reasons, although they are joint benefits. 24Information is not available on how the percentage was determined for this project. However, according to John Okay, economist with the SCS Planning Party in Michigan, in an interview on May 15, 1970, these deductions for FWDRB attributed to land treatment are now in the 5-8% range for Michigan PL 566 projects. They are based on estimated changes in the hydrological runoff curve number (CN), but they are not traced through the FWDRB estimation process. (See ch. 5 on CN-FWDRB sensitivity). The conservation practices in question are less likely to be in the project-benefited lowland area of the watershed (where flooding and impaired drainage are problems) than in the upland area. With respect to computing a benefit cost ratio for them, it should be noted that in addition to the offsite FWDRB attributed to them, there are onsite benefits that include improved soil infiltration and water- holding capacity. As to costs for such a benefit cost ratio, only , some of the costs listed in the PL 566 project work plan table could preperly be included; such a selection would have to be based on a knowledge of the watershed. As a matter of policy, SCS does not compute a benefit cost ratio for land treatment practices. 83 Data Ipputs, Sources and Assumptions Given the SCS procedures for estimating benefits, as systematized into a model here, data inputs are needed to evaluate the project investment. Some data inputs or the methods of obtaining them are specified as a matter of SCS or other Federal policy. A few data inputs are left to the judgment of in-state SCS Planning Party personnel. However, all PL 566 project evaluations are reviewed by a regional—level unit within the agency, known as an Engineering and watershed Planning Unit (E&WPU). This unit reviews various data and procedures that may have been devised locally; also, the in-state SCS Planning Party may request its advice in advance. For purposes of the sensitivity analysis of chapters 5-7, some of these data and procedures will be regarded as assumptions that can be changed. Chapter 4 will provide something of a conceptual backdrop for this sensitivity analysis, although much of the discussion is left to later chapters. Selected variations in crop price, cost, yield and other assumptions that affect the computed farm income are studied in chapter 6, and they are supplemented by variations in farm income directly in chapter 7. The approach of chapter 6 is detailed, like that of chapter 5, and both.of these chapters use the model of SCS procedures for estimating agricultural benefits as develOped in this chapter. Chapter 5 involves a detailed study of the dependence of FWDRB on hydrological assumptions (meaning SCS data and procedures). Chapter 7 is less detailed and takes basic farm income, structural and associated capital cost, and other data as computational model inputs. To avoid duplication, several topics on data and procedures will be deferred to chapters 5-7, except for some comments on data sources. 84 As explained more completely in chapter 6, crop prices and cost adjustment factors are based on USDA projections, either the PLT (projected long term) set prior to 1967, or the AN (adjusted normalized) set since 1967. Crop input-output ratios are based on various sources, including agricultural colleges and experiment stations, local (in-country) SCS personnel, and growers (for some specialty crops). A list of practices is prepared for each crOp and for both input levels. Costs are summarized into two mutually exclusive groups, per unit of output (AVC) and per acre (AFC, non-AVC) costs. As a point of clarification, total production cost per acre = AVC (average variable cost per unit of output) x Y (yield per acre) + AFC (non-AVG, per acre costs). There are two sets of total production costs for each crop, one for each output level.25 Judging by SCS practice in Michigan, per-acre AFC costs for the two output levels (with and without project output levels) are based on present technology with adjustments in certain factor-use rates to achieve the projected (futuristic) output levels assumed. That is, both the with and the without project per-acre AFC costs are based on projected (futuristic) output levels. As explained in chapter 6, base year crop costs are adjusted to conform to the projected crop price levels, but these base year costs ‘r vr 25Costs do not include returns to land or management. The items included in AVG and AFC costs are indicated for the example project in Vondruska, p. 54. However, the categorization of certain cost items may vary among projects and even among cr0ps for a given project, depending on the form in which.the data is available. 85 are developed by pricing the list of inputs just discussed. Input prices are obtained from USDA, the state's_agricu1tural college and experiment station, local suppliers, and farmers.26 When the example project was evaluated by SCS (in 1961—62), crop yields for the two (with and without project) output levels were based on an estimate of what was believed to be possible at the time. As shown in Appendix, Table 6, without-project crop yields for this project correspond quite well with Michigan, 1959-63 state-average yields, with some exceptions. Since that time increasingly futuristic yields have been used in FL 566 project evaluations in Michigan. In 1968-70 SCS used yields for the mid-evaluation—period year. These yields are based on projections by Michigan State University (College of Agriculture) and USDA (SCS, ERS and other) in a joint effort. These yield projections are for various soil management groups, under both drained and impaired drainage conditions, and for three dates (1980, 2000 and 2020). Annual, per-acre yield and cost data are adequate to estimate enhancement benefits or overall net income change benefits, but monthly data are necessary to estimate agricultural FWDRB. SCS uses per unit 26Input price data are contained in: USDA, Agricultural Statistics (Washington, D.C.: USGPO, various years). Also, see W. A. Tinsley, Rates for Custom WOrk in Michigan, Extension Bulletin E-458, revised (East Lansing, Michigan: Mich. State Univ., Cooperative Extension Service, Feb. 1967). 86 of output costs (AVC) directly. The per—acre costs (AFC, non—AVG) are divided on a by-month, by-practice basis. Information on planting and harvesting dates, and late-planting yield reductions is necessary. Further information is needed on by-month portions of the crop destroyed (late season, when replanting would not occur) or portions that must be replanted (earlier in the growing season). Because of the amount of computational detail in estimating FWDRB, the SCS economist may depend on previously developed data, such as for another watershed evaluation or for the region (as in the case of some field crop for the example project).27 Cropping patterns (R) are developed into several sets for each economic reach, one for FWDRB, another for MILUB and a third for LUCB, but they based on two, farm-survey sets. One farm survey set is for the without project condition, usually represented by present cropping patterns in the reach. The second is for the with project condition, and it is represented by present farmers' intentions, given the degree of protection afforded by the project; thus it differs by economic reach as a function of land and human factors. Net Project Effects In this section it will be most convenient to treat FWDRB as being unseparated, that is, as if they were counted as part of EB farm income, 27Percentage of creps harvested by 1/2 month_intervals, monthly FDM values for several yield levels, and depth destruction factors, all for the southern and northern Cornbelt regions, are given in USDA, SCS, Engineering and Watershed Planning Unit,.Milwaukee, Wisconsin, memorandum no. 3 (revised), subject, "Evaluating Floodwater Damages to Crops and Pasture," dated October 23, 1958. Depth destruction factors are available in two sets, either with "duration" (inundation lasts more than 24-30 hours), or without "duration" (inundation lasts less then 24—30 hours); they are based on SCS research. 87 as is done in one sensitivity analysis variation from the SCS model in chapter 7. In assessing SCS assumptions used to estimate net project effects, it must be recognized that some related questions about definitions exist, but they will be deferred, since the concern here is with project-credited farm income. For simplicity the SCS model begins with two projected rates (levels) of farm income, input, output and cost. The shift in farm income is made possible by the complementagy: (l) mainstream project works (structural capital cost or SCK investments), (2) on-farm and inter-farm works (associated capital cost or ACK investments), and (3) management choices which intensify production. Farm income rates are not projected for all individual years in the evaluation period (t = l, ..., T, where T = 50 years or T = 100 years). Adjustments of the without-to-with project income difference to reflect partial and delayed achievement during the EB growth period are discussed in chapter 7, but they do not change the underlying SCS assumption of two income levels. In reality farm income and related rates of input, output and cost change through time. It would be conceptually possible, but much more difficult to specify values for these variables for all years in the evaluation period (t = 1, ..., T). SCS simplifies by picking the rates for one point in time. In the older of the 12 SCS evaluations studied in chapter 7 the selected point in time is closer to project—planning time, but it is closer to the mid year of the evaluation period for more recent SCS analyses. This change in assumption is discussed in chapter 8. Since the two levels of farm income, input, output and cost are for a single point in time they do not reflect changes in underlying 88 general technology. As a point of clarification, Figures 7.3 and 7.4 in chapter 7 may suggest a gradually-increasing withrproject rate of farm income in contrast with a constant without—project rate of farm income, but this gradual increase has to do with the three complementary project effects, as enumerated in the preceding paragraph. Once these effects are completed, with-project farm income accrues at a constant rate . Summary The basically flexible SCS model requires the use of assumed and measured data inputs. The model yields a benefit cost ratio, which is used to justify the investment economically. The focus of this chapter is on the computation of agricultural benefits, which are the main thrust of the small watershed program. Agricultural primary benefits (not counting secondary and redevelop- ment benefits) consist essentially of the difference in net income for the aggregate of crop enterprises in the watershed benefited area for with and without project conditions. Projected yields and prices are assumed, and present-technology input combinations are adjusted to provide the projected yield levels. Farmer-indicated output combinations (cropping patterns) are assumed. The SCS model does not optimize these input and output combinations to maximize profits for the farm unit, nor to maximize benefits for the watershed or economic reach. For policy reasons, SCS estimates separate Categories of agricul- tural benefits: FWDRB (floodwater damage reduction benefits), for the river overflow Zone; MIL B (more intensive land use benefits), for already cropped farm land; and LUCB (land use change benefits), for previously 89 uncropped farm land (not for land classified otherwise).. Special attention is given to the hydrological and economic aspects of FWDRB. The other two categories, LUCB and MILUB, are simpler to compute and explain. CHAPTER IV SOME CONCEPTUAL PROBLEMS This chapter is intended to provide some conceptual background for the sensitivity analysis of chapters 5-7. The topics include: efficiency criterion rules, social discount rates, and sensitivity analysis--a prologue. The purpose of the sensitivity analysis of chapters 5—7 is to study the benefit and efficiency-criterion yardstick responsiveness to changes in underlying hydrological, crop enterprise and other variables and procedures. If one cares to question the agency's assumed values for different variables, many of the assumed values can be changed with this technique. Some economists have focused on what they believed to be key variables, changed the values used, and provided some measure of possible "optimism bias" on the part of the agency. Different views on this matter are discussed under the topic of social discount rates; that is, whether discount rates or benefits and costs or both should be corrected. It would seem that both may require correction, if one views agency estimates as being biased, and if in addition the discount rates do not properly reflect one's concept of the social discount rate. One explanation for the view that agencies provide biased data is that they are concerned with their own survival. Another is that efficiency-criterion yardsticks pay too much attention to the matter of additions to national income, and not enough to the matter of its 90 91 distribution. Therefore, it is argued that agencies purposefully choose procedures and data to enhance a project's worth as measured by efficiency-criterion yardsticks to compensate for ignored effects in redistributing income. In the first section of this chapter the matter of efficiency- criterion decision rules or yardsticks is taken up. The discussion of efficiency-criterion yardsticks, the net present value, internal rate of return and benefit cost ratio, often focuses on their useful— ness as ranking devices. While ranking is de-emphasized in chapter 7, where 12 Michigan PL 566 projects are studied, the measuring quality of these devices is employed. The matter of budget constraints is also taken up in the discussion of efficiency criteria. Secondly, the question of social discount rates is explored. Clearly, discount rates are key variables in decision-making models. However, the importance of the topic extends beyond the question of discount rates per se, for economists have used it as a vehicle for discussing risk and uncertainty, rates of return on investment, capital formation, the division of investment between public and private sectors, the allocation of resources between short and long term projects, and the provision for the "future." The whole discussion provides a convenient rationale for studying the sensitivity of investment criteria data to changes in discount rates, as well as other underlying variables, as is done in chapters 5-7. Efficiency Criterion Rules For independent projects, in the absence of budgetary or other constraints, A. R. Prest and Ralph Turvey summarize four investment 92 decision rules which are based on the concept of economic efficiency, meaning the maximization of the present value of benefits less costs: (1) The NPV (net present value) rule: select all projects where the NPV is positive, that is where the present value of benefits exceeds the present value of costs at the chosen discount rate. (2) The B/C rule: select all projects where the ratio of the present value of benefits to the present value of costs exceeds unity at the chosen discount rate. (3) The IRR rule: select all projects where the internal rate of return exceeds the chosen rate of discount. (4) The constant annuity rule: select all projects where ‘ the constant annuity with the same present value as benefits exceeds the constant annuity (of the same duration) with the same present value as costs at the chosen discount rate. Providing benefits and costs are defined consistently, the IRR equals the chosen rate of discount when the BIG ratio equals one and the NPV equals zero. Prest and Turvey indicate that if the chosen rule is not algebraically the equivalent of these four, either error or a different maximand is involved. The SCS procedural variations from the closest of these four rules, the BIO ratio, will be discussed in chapter 7. Ranking Divergences Strictly speaking, the alternative decision rules just expressed are border conditions, meaning that they may be used to select or reject projects. Furthermore, these decision rules apply to independent (not interdependent) projects in the absence of budget constraints. These restrictions may also affect ranking. Does a 1A. R. Prest and R. Turvey, "Cost-Benefit Analysis: A Survey," The Eponomic Journal, vol. 65, no. 300, December 1965, pp. 683-735, esp. pp. 703-704. unlrv jun mug 3H1 nvuhw :J anJ .9731} no mi /QE 9! aywzq Elm jwwlna :uiu1 VJIHHHL Jan 3); UJUQC W. tuf: / JUUflflffl? 9:3 U7 KUJJIJ;LQ;;UJZ 5r: :7‘4 1 ~‘ I . I -' . ' r‘l ' - " ' u ‘ .l' f l h" ' :>,l.u .;.h1 l.u!L/ :4...,n .'H/ ..u .L. (1) (1“! UL'IJ ,_ r J ‘).r1_(J(J .ani A J .' . -4, . —. 7' v - ', - v ;. FVK :‘I r- . 913 J. a.un3 .w UHILJ JHQMUYfl UMJ ‘JUQJUU 831 winds v.4wiu1q [1L Juann zuiuw .3[ on. (L) 'w'luv H.wr'xw; u!:} a] :‘Jii'mr7d I!) ujsz jrrjnuvc; uflj .ujtrz Jnxur5910 rriaono éxfj 11; vjlruJ abssuxis w‘zwiu: ;'}')'.{'1)'1l! 111'. 1093-13: 10111] 1.)". {mil (4) 'rnjr)u.;‘i) ‘11) '93:;1 r15zn:: o ur.J :.;y)<;9; HomV xn woucoaoammow ”Aooma kum< .Hfioasoo ugh u.o.n .aouwafismm3v . . . mwumvomum oofium afiuouoH .Haocsoo woousowom “mama .m.D "vmumamu pom z< Mom .Ammma .«nm: ".o.a .aouwcflnmmzv . . . maofiuoofioum umoo was moaum Hmuouaooauw¢,.mz< was mm¢ .accmu cccaum AanmN v NNNV omN mmN msN oqN mmN Nam .cc>chmu mmoaum ooN we wrap «N-ONmN .m .roNz can: more .moofiuoonoun wmumafiv< ucouuso .m.= .qu sopcmum moofium,vmufiamauoz concoma .mumn ooaum mono omuomaom .o.o manna 199 in the Water Resources Council in 1966. The shift from PLT to AN prices would adversely affect projects with benefits dominated by grains, sugar beets and cotton, because all of these crops suffered price reductions. These cr0ps are directly affected by government programs. Therefore, to obtain their AN prices from their current normalized prices, a slight downward reduction was applied.11 The current normalized prices are approximately the same as the USDA average for the United States for 1960-64. The PLT sugar beet price includes Sugar Act payments, whereas the AN sugar beet price does not (see Appendix, Table 7). Further adjustments are necessary to obtain prices for particular states from the PLT and AN prices for the United States. The Water Resource Couuncil's alteration of 1960-64 prices to obtain the AN prices did not remove the full effect of government programs, as suggested by George Brandow's projections for 1965. For grains the AN prices are roughly halfway between the PLT prices and Brandow's. For the sensitivity analysis the author reduced 1959-63 11United States, Water Resources Council, Interdepartmental Staff Committee, Interim Price Standards for Planning and Evaluating Water and Land Resources (Washington, D.C.: The Council, April 1966), p. 2: The adjustments in normalized prices to reduce the influence of Government programs are intended to reduce or remove most direct price support effects or payments under such programs rather than the full effects of all production adjustment programs. For crops seriously in surplus, further constraints in the form of addi— tional price adjustments or acreage limitations may need to be applied. Restrictions on surplus crop production as a primary, dominant source of project benefits are discussed in chapter 2. 200 prices by 20% to complete the Brandow series (see data in Appendix, Table 7).12 Crop Cost Data Adjustments As explained in chapter 3, costs for various practices and inputs are obtained by the SCS in-state Planning Party from USDA, state agri- cultural colleges and experiment stations, local suppliers and other sources. SCS uses a crop enterprise rather than farm budget approach. Essentially, a list of inputs and practices is priced for each enter- prise for a base year. Then the costs are adjusted downward using a factor that is intended to put crop costs on the same projected basis as the crop prices used in evaluating the watershed investment. The factor is obtained by dividing the proper projected value of the USDA index of prices paid by farmers by the value in the base year. For the PLT factors the all-items index is used, but for the AN factors the production-items index is used.13 12Walter Wilcox, "Agriculture's Income and Adjustment Problems," in U.S. Congress, Joint Economics Committee, Economic Policies for “Agriculture in the 1960's, 86th Congress, 2d Session (Washington, D.C.: 'USGPO, 1960). The prices were developed by George Brandow using a de— 'mand model for 1965, assuming discontinuance of Federal surplus diSposal and storage programs; plus international stability; and upward trends in ‘population, productivity and real income per capita. The USDA's index of prices received by farmers was projected to decline 21% from 1959 using these assumptions. 13As shown in Table 6.6, the PLT value for the USDA all—items, ‘prices paid by farmers index is 265, and the AN value, for the produc- ”tion-items index is 272. Using the reapective index values from USDA, IERS, Demand and Price Situation, DPS-115 (Washington, D.C.: USDA, Feb- :ruary 1968), the author computed the following factors, which may differ frrnn those used by SCS for some years, apparently because preliminary idata.were used. PLT factors are first: 1960, .88, 1.09; 1961, .88, 1.02; 1962,..86, 1.01; 1963, .85, 1.00; 1964, .85, 1.01; 1965, .83, .99; .1966, .79, .95; 1967, .77, .95. PLT factors and prices were last used 201 Depending on what happens through time, the cost adjustment factors may either compensate for rising production costs (if the costs and the index actually move together), or decrease the costs used in the evalu- ation (if costs remain relatively constant, while the index rises). SCS also applies these cost adjustment factors to project operation-maintenance costs for both associated works (cost item ACOM) and structural works (cost item SCOM). However, the author did not adjust these costs. Alternative Crop Yields While many variables affect the project-credited farm income, it is relevant to emphasize the effect of crop yields, because of the compara- tively high degree of benefit responsiveness to slight changes in yield assumptions, as shown previously in Tables 6.2-6.4. Yield assumption changes on a by-crop basis are more problematic than those for prices, because yields vary with soil type, climate, water problems, and other factors peculiar to the watershed. While state average yields probably do not represent watershed conditions sufficiently well for SCS usage in project evaluations, 1959-63 state average yields for Michigan correspond surprisingly well with the without-project (Y1) yields used by SCS for the Mill Creek project, with the exception of vegetable crops, as shown in Appendix, Table 6. USDA's 1959-63 state average yields for Michigan were adjusted to form the necessary without (Y1) and with project (Y2) yield sets to use in the SCS benefit computation model of chapter 3. fl 13by SCS in 1966. See Appendix, Table 1, for factors used by SCS for 12 Michigan PL 566 projects. Also, see footnote 8 on conversions used by the author. 202 Table 6.7. Effects of Alternative Crop Prices, Costs and Yields. ASsumption ' ‘ Benefit component index .. w .M... ..,_... change , FWDRB .MILUB LUCB ‘ Overall . B/Clratio' Base, value $75,715 $130,362 $74,604 $282,065 6.96 Base, index 1.00 1.00 1.00 1.00 - — (A) Brandow (no government support) prices, costs for 1960a,c 0.61 0.47 -0.05 0.37 2.60 (B) State yields (Y1: “25%; Y2, +25%), costs for 1960b,C 0.58 2.05 0.44 1.22 3.53 (C) Prices, costs and yieldsaab,C 0.44 1.07 -0.11 0.59 4.07 Source: Approximations of SCS estimate for the North Branch of Mill Creek watershed and sensitivity analysis variations thereof. FWDRB are for cr0ps only, but overall benefits include minor non-crop FWDRB. aPrices based on projections by George Brandow for 1965, without government programs; see data in Appendix, Table 7. bYields, based on Michigan state average for 1959-63 (shown in Appendix, Table 6), plus 25% for Y2 and minus 25% for Y1, with proportional per-acre AFC cost adjustments for MILUB and LUCB only. c1960 production costs: SCS PLT costs x 1.1236, see main text, footnote 8. Tables 6.7 and 6.8 can be used to show yield change effects. Alternative B in Table 6.7 shows the effect of using adjusted state average yields for 1959-63, 1960 costs and the original SCS PLT prices, while alternatives A—E, Table 6.8, are similar except that 1959-63 state average prices are used. These 1959—63 prices provide benefit index Values quite close to the base value when used in combination w1111.l960 costs (compare.the base index and altnerative A in Table 6.8); therefore, for present purposes this assumption combination will be. regarded as. an approximate equivalent of the. original SCS assumptions. ...-l4 203 While it becomes something of a guessing game, a little discussion of which alternative assumptions would approximate the SCS results may be useful. In terms of overall benefits, adjusting 1959-63 state average yields to -25% (for Y1) and +25% (for Y2) overshoots the orig- inal SCS estimate (see alternative B, Table 6.7). Adjustments of -25% (Yl) and 100% (no adjustment for Y2) are about right (see alternative B, Table 6.8). However, yield set 1 would have to be raised above the state average yields to approximate the SCS estimate for FWDRB, and yield set 2 should be raised to bring overall benefits up to the level of the SCS estimate. State yields adjusted to -10% (Y1) and +10% (Y2) better approximate the original SCS estimate of MILUB, only slightly improve the FWDRB approximation, and leave overall benefits lower (see alternative C, Table 6.8). Clearly, it is possible to effect a considerable range of overall benefits. Any of a number of alternative yield assumptions would leave the NMBC project with a benefit cost ratio well above 1:1. Because of the policy constraint on MILUB and LUCB, effects of dif- ferent yield assumption alternatives are interesting. For example, alternative D, Table 6.8, reduces MILUB (to 54% of the base level) and LUCB (to 43% of the base level), but does not greatly affect FWDRB (84% of the base level), and still provides a benefit cost ratio of 4.14/1. SCS Yielg Assumptipns SCS employs yields for the mid—evaluationrperiod (some 25 or 30 years. in the future from the date of planning), judging by; SCS practice in Michigan, 1968-70. The yields themselves are the result of joint 204 agricultural college and USDA efforts. The yields are for various soil management groups, under wet and well-drained conditions,with no specific reference to flood problems in the assumptions of projection. For the example project (evaluated by SCS in 1961 and early 1962), the yields were based on what was believed to be possible at planning time. Thus, as indicated in chapter 3 and further discussed in chapter 8, SCS yield assumptions represent a shift from what is technologically and otherwise possible in the near future to what is possible in the more distant future, that is, for both with and without project conditions. With respect to inputs and input-output ratios, SCS adjusts fertilizer, seed, chemical spray and possibly other input rates for the higher yield levels to obtain the per-acre cost data (AFC data). The effect of using higher yields (in both with and without project yield sets) would be to increase overall benefits and to emphasize FWDRB, judging by the yield sensitivity analysis for the Thus, increasing yield set 1 boosts FWDRB; yield set 2 example projects. Simultaneously raising both yield sets 1 and 2 similarly affects LUCB. increases MILUB. The overall percentage increase in benefits would at least equal that in yields, and could be two or three times as much. Conceptualizing the effect of using higher yields to evaluate PL 566 projects requires some appreciation of the SCS-assumed growth patterns. for projectrcredited farm income. FWDRB accrue at their full 100% rate (based on without-project yields and per—acre costs) at time zero in the evaluation period (usually 50 years). However, enhancement benefits. (EB) are based on gradual and delayed achievement of the move from without to with project farm income. Despite the use of 15—20 year development periods for the achievement of with-project income levels, 205 the simplifying SCS assumption of instant installation means that 1/2 to 3/4 of the average annual benefit rate accrues at time zero of the evaluation period, taking into account both EB and ‘.F'WDRB.14 Because this proportion is so high, and because project benefits are increased by yield increases, it is relevant to question the use of mid- evaluation period yields. The use of mid-evaluation period yields may be viewed as a proxy for computing yield and farm income data on a year by year basis. Assuming that yields would increase over time, early-year yields are overstated and later-year yields understated by the use of mid-evaluation The net effect is an overstatement of project-credited period yields. farm income in terms of present values. This will be reconsidered in chapter 8. Multiple Assumption Changes While previously considered assumption changes were for one variable at a time, Tables 6.7 and 6.8 show the effects of combined price-cost, yield and cropping pattern changes for the Mill Creek project. For all alternatives shown, crOp costs are restored to their 1960 level, as estimated by SCS; that is, the SCS reduction to PLT cost levels has been removed. The USDA state average prices for 14SCS--assumed growth patterns for project-credited farm income are discussed in more detail in chapter 7. The proportions cited are from Appendix, Table 3, last column; they are computed as follows: (benefits, time zero) / (average annual benefits), using the SCS defi- nition of benefits in which associated costs are deducted. This is a proxy for the proportion: (project-credited farm income, year zero) / (eventually-achieved farm income). The definition of benefits includes OTHERB (see chapter 7), but excludes secondary benefits; it is not the equivalent of net cash flow, for SCK and SCOM are omitted (see chapter 7) . 206 Michigan, 1959-63, include the effects of govermnent farm programs, but the prices based on George Brandow's projections for 1965 Specifically exclude the effects of government farm programs. The l959v-63 state average yields for Michigan are used with several adjustment factors, constant for all crops, and differing for the with and without project yield sets. There are four pairs of yield adjustment factors (I, II, III and IV, as defined in Table 6.8) and their use is repeated in Table 6.8 for alternatives B-E, G-J and M-P. Finally, cropping patterns for economic reach 3 are applied to the other two reaches, thereby transferring benefit dependence from vegetable to field crops, which are perhaps more typical of Michigan agriculture. In Table 6.8 each new assumption change is shown alone (alternatives A, F, K and L) and then combined. The shift from PLT prices and costs (base index) to 1959-63 state average prices and 1960 costs (alternative A, Table 6.8) did not greatly affect any category of benefits. Both sets of prices include the effects of government programs. To exclude the effects of government programs, prices based on George Brandow's projections were introduced, and all categories of benefits declined significantly (Table 6.8, alternative F). Introducing adjusted 1959-63 state average yields produced varying effects. Progressing through yield-adjustment assumptions I-IV, FWDRB improve, because yield set 1 increases; MILUB and LUCB decrease, because yield set 1 increases and yield set 2 de- creases (alternatives B-E and G-J, Table 6.8). If the negative LUCB benefits were deleted for the most severe of these alternatives (.1 in Table 6.8), the project would have a benefit cost ratio far below the base level of 6.96/1. Nevertheless, FWDRB 207 would represent about 78% of average annual benefits compared to 27% for the base estimate data. Therefore, altering yield and price assumptions can be viewed as a means of changing the apparent emphasis on various kinds of benefits (see chapter 2). Flood prevention benefits (FPB) FWDRB $39,817 (+$1,383, non—crop) FPB-MILUB A 5,750 Sub-total, FPB $46,950 Agricultural water man- agement benefits (AWMB) 5,750 Total benefits, costs, B/C $52,700 / $40,520 = 1.30 Introducing economic reach 3 cropping patterns reduces the benefit cost ratio to 1.24/1 (alternative K, Table 6.8), quite close to the base estimate for reach 3 (B/C = 1.22/l), and this demonstrates the impor- tance of vegetable crops (including potatoes) in producing the base level benefit cost ratio (6.96/l) for the Mill Creek project. Even modest price reductions or cost increases would be sufficient to reduce the benefit cost ratio below unity for a project with cropping patterns like those of reach 3, as shown in Table 6.5 (alternatives with super- script a after B/C ratio). For example, restoration of crop costs to the 1960 level, or the use of 1959—63 state average prices with PLT costs, or 1959963 prices and costs, or the more recent SCS AN prices 2nd costs. Assuming reach 3 cropping patterns, 1960 crop costs, and crop prices at remove the effect of government programs reduces the benefit cost tio to $0.50/1. Beyond this, progressing through yield adjustment ctors I—IV for 1959563 state average yields merely worsens the ratio lternatives M-z-P, Table 6.8). 208 Table 6.8. Effects of Alternative Prices, Costs, Yields, and Cropping Patterns. ' ‘ Assumption ‘ Benefit sensitivity indexes changes FWDRB MILUB ’ LUCB 1 ' Oyerall B/C ratio Base, value $75,715 $130,362 $74,604 $282,065 6.96 Base, index 1.00 1.00 1.00 1.00 - - l959v-63 Michigan State average prices (P), 1960 costs (C), and yield (Y) variationsa. (A) P 5 C only 1.07 1.07 0.90 1.02 7.13 (B) Y-I 0.70 1.40 0.47 0.96 6.71 (C) Y-II 0.81 0.98 0.49 0.80 5.59 (D) Y-III 0.84 0.54 0.43 0.59 4.14 (E) Y-IV 0.88 0.09 0.37 0.38 2.63 Projected 1965 prices (P), 1960 costs (C), and yield (Y) variationsbaC (F) P & C only 0.61 0.47 —0.05 0.37 2.60 (G) Y—I 0.44 0.85 -O.15 0.48 3.32 (H) Y-II 0.49 0.53 -0.18 0.33 2.23 (I) Y-III 0.51 0.32 -0.22 0.23 1.60 (J) Y-IV 0.53 0.09 -0.25 0.12 0.84 Projected 1965 prices (P), 1960 costs (C), economic reach 3 cropping patterns (R3), and yield (Y) variationsbfi (K) R3 only 0.18 0.10 0.30 0.18 1.24 (L) R3,P&C only 0.11 -0.15 -0.14 -0.07 -O.50 (M) Y-I 0.09 -0.44 -O.30 -O.26 -l.78 (N) Y—II 0.10 -0.50 -0.30 -O.28 -l.94 (0) Y-III 0.11 -0.53 -0.30 -0.29 -2.04 (P) Y-IV 0.11 —O.58 -O.31 -O.32 -2.22 Source: Approximations of SCS estimates for the North Branch of Mill Creek watershed and sensitivity analysis variations thereof. FWDRB are for crops only, but overall benefits include some minor non-crop FWDRB. aAlternative A only uses 1959-63 costs (SCS PLT costs x 1.1363) rather than 1960 costs as stated (SCS PLT costs x 1.1236). See prices, Appendix, Table 7. 1)Alternative yield sets are based on 1959v-63 Michigan State average 'ields (Y-M), as. shown in the Appendix, Table 6, adjusted as follows to btain without-PPIOJect (Y1) and with-eproject (Yz) yields: (I) 11, = Y-M x 0.75 and Y2 s Y—Mx 1.00; (II) Y1. a Y-M x 0.90 and Y2, = YHM x 1.10; (III) Y1_- Y-M x 0.95 and Y2 =- Y-M x 1.05; and (-IV) Y1_= Y-M x 1.00 and Y2, = Y—M x 1.00. r—acre AFC costs adjusted proportionately for MILUB and LUCB only. CPrices based on projections by George Brandow for 1965, without vernment programs; see data, Appendix, Table 7. 209 Alternative Cropping_Patterns Cropping pattern data are critical in determining project benefits. The use of reach 3 crOpping patterns. for all reaches of the Mill Creek project would reduce the benefit cost ratio from the base level of 6.96/1 to 1.24/1, as shown in Table 6.8 (alternative K). The effect of altering cropping patterns may be suggested by the data in Table 6.9. When cropping patterns and by-crop loss values are multiplied and summed for all crops they provide the composite acre values (CAV, typical acre loss values for the floodplain), which are used in conjunction with acreage flooded data to determine floodwater damages (FWD) and reduction benefits (FWDRB), as explained in chapter 3. The differences between composite acre values for reaches 1, 2, and 3 are due to crOpping pattern differences (Table 6.9). Depth of inunda- tion effects are by comparison insignificant.15 CrOpping intensifica- tion to the with-project economic level16 is less important than broader 15Tolley and Freund similarly conclude that the type of agriculture represented here by crOpping patterns is quite important. However, the author is more inclined to reviewer Cohee's comment that probabilities of error associated with hydrological data should not be lightly dis- missed; Tolley and Freund assign probabilities to different variables, and indicate that those associated with the type of agriculture require the most judgment, whereas those dealing with hydrology are more objec- tive. This may be, but FWDRB are extremely sensitive to changes in hydrological assumptions, as shown in chapter 5. See George S. Tolley and Ralph. Freund, Jr., "Does the State of the Data Suggest a Program for Modifying Planning and Evaluation Procedures?" with comment by Melville H. Cohee, in G. S. Tolley and F. E. Riggs, Economics of Watershed Planning (Ames, Iowa: The Iowa State University Press, 1961), pp. 127-14]. 16CAV for the intensified cropping situation assume Y2, per acre costs based on AFCZ, and withrproject cropping patterns based on 100% completion and achievement of the MILUB and LUCB initial net return levels. 210 .wsfivasou mo monsoon Huuou ou moo uoa has mmn3h you massaoo ca sumo muoomoue xoouo Hafiz onu mom nonmafiumo mom mo aofiuwafixouoo< “mousom No.N~ na.om wm.mm «n.0H no.0q oo.ow «N.HH Ho.om n~.mo ow.w m me.amm no.mom mumaaop .aoHumosan mo mason was Ho>oa useuso .nomou kn .mmsam> ouoo oufimomaou N eueuu .N Hu>uH H recur .N HueuH N recur .H Hu>uH H eueue .H HceuH No.ooH No.ocH No.ooH Ne.mN Nm.ew Ne. Hm Huuoa no- H.N H.H -o- m.H e. H ee.ee .uua euuuu .uuueaeuso -o- no- e.H no- no- o. H «N.om .uea euuum .uusuuuH -o- m.e m.H no- e.N e. H mH.cm .uea euuuu .umueeuu no- uou a.m -o- lo- a. N ee.mHN .uea euuuu .auuHuo uo- m.eH o.NN Io- m.e m. NH mm.moH .uee euuum .uuouuuo no- .0. m.e no- .o. s. N No.oNH .uea euuuu .ueoHco no- m.em H.mm no- N.HH m. mN HH.ea uuouuuoe no- no- H.o no: no- a. e mm.o ecu uuuum eeuH H.H «.0 -o- m.m N.a no- ee.H measure unusuauuu -o- m.H m.H no- m.N m.HH mm.N are uuHuuHa a.m e.o no- m.eH o.c no- wH.m are uu>oHo -o- e.cH N.NH no: m.m m.HH NN.¢HH uuuue numem a.em m.H no- m.eH H.m no- Hm.mH mucus s>uz m.cH e.N N.H H.w H.c m.e mN.HH uuuuas .uaurz N.m a.o H.o H.w o.m e.H mc.m eHuum .uuuo H.MH m.N s.o o.m o.m e.H oo.mN umaHHu .euoo No.mN Ne.mH Na.N Nm.HH Na.NH NN.m mm.NN eHuum .euoo m mommy N momma H mono“ m comma N mommy H momma w flmmoa some mono onulwn vousnfiuucoo (mono may ou mouemam .H nuoom mmnzh mo owmuaoouom sHmHepoon saw no owmucoouom .H Ho>oa .ouomluom .mmnzm ou muoausnwuuaoo mono new mduouuom moaneono .m.o manna 211 changes in the type of agriculture, represented by the shift from predominantly vegetable crop and sugar beet farming (reach 1 data, Table 6.9) to field crop farming (reach 3 data). Cropping patterns are expressed in Table 6.9 as percentages of the floodplain acreage devoted to each crop, but differ from those indi- cating the preportion of FWDRB associated with each crop. For reach 1, 3/4 of the FWDRB, but only 1/2 of the acreage is represented by sugar beets, potatoes and carrots. These crops account for 1/2 of the FWDRB and 1/4 of the area for reach 2. Going to reach 3, navy beans and clover hay occupy the same acreage, but navy beans account for over 1/3 of the FWDRB, and clover hay, about 6%. Winter wheat is grown in all 3 reaches, occupying 6.5%, 6.1% and 8.1% of the respective areas, and it accounts for 1.2%, 2.4% and 10.8% of the FWDRB for reaches 1, 2 and 3.17 Summary The three categories of crop benefits for PL 566 projects in agricultural watersheds, FWDRB, MILUB and LUCB, differ in sensitivity to changes in crop price, cost and yield data. Increments or decrements for each of these variables were introduced into the SCS model as constants for all crops, for the range 10% to 50%, by units of 10%. Benefit responses remained constant over the considered range, with coefficients of response varying from one to about eight. FWDRB are . —-— v f 17The cropping pattern and benefit percentages differ for FWDRB, LMILUB, LUCB and overall benefits. -Ragarding the several cropping pate tern sets, see John Vondruska, Estimating Small Watershed Project Benefits: A Computer Systematization of SCS Procedures (East Lansing, Machigan: Department of _Agricultural Economics, Michigan State Univer— sity, February 1969), pp. 58—59 (discussion of R). 212 generally least sensitive, whereas MILUB and LUCB are extremely sensi- tive to yield changes. The response of MILUB to simultaneous yield set 1 and 2 changes is far less, ranking with those for prices and costs. Judging by the Mill Creek project, PLT (projected long term), AN (adjusted normalized) and Mdchigan state-average (1959-63) prices and the respective crop-cost adjustments provide about the same level of benefits. Using 1964-66 state-average prices and the PLT crop-cost level virtually doubled benefits, but benefits were considerably reduced when the comparable 1964-66 cost adjustment factor was applied. 0n the other hand, using a price set based on projections (by George Brandow, for 1965) to remove the effect of government programs and 1960 crop costs reduced the benefit cost ratio to 2.60/1 (about 1/3 of the base ratio of 6.96/1). Because project benefits are quite sensitive to crop yield changes, it is relevant to consider the assumptions underlying the yield data. As will be explained in chapter 7, the with-project level of farm income is achieved quite rapidly in the evaluation period (see Figures 7.3 and 7.4). According to SCS practice in Michigan (in the late 1960's), mid- evaluation-period yields are used, but they would appear to result in overstated benefits, as will be discussed in chapter 8. Project benefits are also sensitive to cr0pping-pattern assumption changes. The combined effects of adverse assumption changes of the type considered in this chapter are sufficient to reduce even a rather high. ‘benefit cost ratio to unity or below. The assumptions studied in this chapter affect the level of projectrcredited net farm income and they 213 will be taken up again at the close of chapter 7, as part of a study of 12 Michigan PL 566 projects. llII ...I.|¢nl.l‘ll‘|lllv j CHAPTER VII INVESTMENT CRITERIA: BENEFIT AND COST TIMING, PATTERN AND LEVEL ASSUMPTIONS Twelve Michigan PL 566 project evaluations are studied in this chapter, with emphasis on benefit and cost timing and pattern changes. The sections include the following: methodology and investment criteria, ranking projects, interest rates, enhancement benefits analysis, cash flow timing and instant installation, FWDRB (floodwater damage reduction benefits) redefined, adverse farm income and capital cost changes, and summary. The twelve projects have work plans dating from 1959 to 1968, and are primarily oriented toward assisting farmers to improve their income or asset situations via flood control, drainage and irrigation (Sturgeon only). However, the Sturgeon emphasizes natural resource investments for recreation, as does the more recent Maple River project group, which is not studied here. Detailed data were not available for some projects; hence, they are not studied here. As in chapters 5 and 6, project base estimate data are developed for comparison with sensitivity analysis results. Data for the twelve PL 566 projects were obtained from SCS files, and supplemented by other data from work plans. However, for some projects, SCS prepared subsequent, supplementary work plans and 214 1U" 215 supporting documentation (in SCS files). The projects‘ data are shown in the Appendix, Tables 1—5. Methodology and Investment Criteria In this section consideration will be given to the mathematical formulation of investment criteria used in chapter 7. While some other criteria are studied, only three will be employed throughout the chapter: the SCS B/C (the SCS annual benefit cost ratio), the IRR (the internal rate of return), and the NPV sum (net present value sum) for all 12 Michigan projects. In all cases secondary benefits are excluded.1 The SCS Benefit Cost Ratio (SCS B/C) The SCS B/C computational routine provides numerical benefit cost ratios that are the equivalent of those that would result from use of actual SCS procedures described in chapter 3.2 The SCS annual benefit cost ratio may be simply formulated as SCS B/C = (FWDRB + EB + OTHERB) / (SCK x A4 + SCOM) 1Based on discussions with John Okay, economist with the SCS Planning Party in Michigan, in late 1969, secondary benefits may not be used to justify a project, i.e., to bring its benefit cost ratio above unity, but they may be used to supplement a ratio already above unity. Their use seems ambiguous to the author and subject more to agency directives or non-written understandings than to long-standing policy. The agency‘s policy—procedural guide condones use of nonnnational, regional or local secondary benefits, in compliance with.Senate Document ~97 (of 1962); see USDA, SCS, Watershed Protection Handbook (Hashington, .D. *C., SCS, 1967), secs. 102. 02213 and 102. 02214. V 2Regarding reconciliation of the author's base estimates for chapters 5%] and the SCS agency estimates, see chapter 1 and also foote» 'note 19 in chapter 3. The.two sets of BIG ratios are shown in Appendix, Table 3 and differ usually by less than 5 percent. 216 In words, it is the comparison of annual benefits to annual costs, with mainstream project costs amortized in the denominator of the ratio (SCK x A4). Rather than compute EB (enhancement benefits) as in chapter 3 (following SCS procedures), the author has chosen to compute EB cash flows for all years in the evaluation period (years t = 1, ..., T, where T = 50 or 100). These cash flows differ from those computed for alternative, non-SCS investment criteria in that amortization factors are used. SCS B/C is computed as follows: SCS B/C = (FWDRB + OTHERB + EB) / (SCK x A4 + SCOM), and '1' 3 EB = A3 x {2% if [DNR - (ACK x Al) — (ACKi if if,l t-l x A2) - ACOMi f,2 f Variables are defined as follows: A: an amortization factor; generally used to determine equal annual loan repayments over a period of years including both principal and interest; used by SCS to convert capital or present values into annual equivalents. Interest rates and amortization periods differ by project (see Appendix, Table 2, including footnote b). The use of amortization factors is discussed later in this section. Factor Interest Period Use A1 r1 15-50 yrs. ACKif’l amortization A2 r2 25-50 yrs. ACKif 2 amortization A3 r3 50-100 yrs. EB amortization A4 r4 50-100 yrs. SCK amortization ACK: associated capital costs for on-farm (ACKif 1) and inter-farm (ACKif’z) works of improvement; excludes SCK. ’ ACQM: annual Operationsrmaintenance cost for ACK; excludes SCOM. DNR: difference between with.and without project net returns (net farm income, annual, see chapter 3, including Tables 3.5 and 3.7); exclude FWDRB. EB; enhancement benefits (annual). 217 FWDC: change in withrproject floodwater damages (annual) resulting from computing damages at the without-project level of economic activity and the increase in damages due to intensification of economic activity (see chapter 3, Table 3.7, footnotes 15 and 16, and chapter 6, fOOtnote 4). FWDRB: floodwater damage reduction benefits (annual), the difference between with and without project damages; FWDRB = FWDl — FWDZ. Crop FWDRB are one form of project-credited farm income, exluding DNR. OTHERB: annual, non-FWDRB and non-EB benefits; exclude secondary benefits. P: the proportion of the initially-estimated EB cash flows achieved in any one year; maximum P ranges from 75% to 100%, depending on the project (see Appendix, Table 2, including footnote c). r: an interest rate (see A, amortization factor, preceding, and Appendix, Table 2, for rates used by SCS); as many as four rates may have been used by SCS in evaluating each of the 12 studied projects. SCK: structural capital costs for major mainstream works; exclude ACK. SCOM: annual operations-maintenance costs for SCK; see ACOM. Values used for these variables are shown in Appendix, Tables 1—3. Subscripts are defined as follows: if: EB cash flows (DNR, ACOM, FWDC and amortized ACK) are usually divided into components by SCS, and separate EB achievement rates (P1,if,t) are applied to compute annual cash flows for any year t. For the EB sub-category more intensive land use benefits (MILUB), if I 1. For the EB sub-category land use change benefits (LUCB) associated with land formerly in permanent pasture or idle uses, if = 2. For LUCB associated with.land formerly in farm woods, if = 3. To obtain the sum of all cash flows for any one year (t), it is necessary to sum over the range.if = 1-3. For some projects, SCS used only one or two categories of EB (see Appendix, Table 2, including footnotes org), and in this case the additional variable values were set to equal to zero for purposes of computer programming. 218 t: refers to a year, t = l, ..., T, where T = 50 or 100 years, the evaluation period lengthn. The computation of EB (enhancement benefits) is the nwst complicated operation in obtaining SCS B/C, an annual benefit cost ratio. Prior to discounting, summation over t, and application of the amortization factor A3 in the formulation of EB, annual combined EB cash flows during the evaluation period may be visualized as a polygon as in Figures 7.3 and 7.4. Time is measured along the horizontal axis, and the height of the polygon for any year t is determined by the prOportion P1,if,t° SCS procedures involve slicing the polygon into horizontal segments (see chapter 3, Table 3.8), while the approach used in this chapter involves slicing the polygon into vertical segments, with one vertical segment for each year t (where t = l, ..., T, and where T = 50 or 100 years). In either case the segments are discounted and summed to obtain a present value which is then amortized. According to SCS assumptions the com— bined EB annual cash flows (DNR, ACOM, FWDC and amortized ACK) are a single entity which intially grows after the fashion of a decreasing rate growth curve (as in Figure 7.3), but with linear approximations. Subsequently, the combined EB annual cash flow rate reaches a maximum and becomes constant after the growth period is completed (after 15-30 years, depending on the project). The maximum annual EB cash flow rates are determined by the maximum value of the proportion P1,if,tr and this is in the range of 75 to 100%, depending on the project (as shown in Appendix, Table 2, columns 12 and 16). 219 The SCS instant installation assumption3 complicates the matter in that all annual cash flows are advanced five years with.respect to EB. Thus, the linear approximation of a growth curve in Figure 7.3 and repeated in Figure 7.4a becomes the linear approximation shown in Figure 7.4b, as will be discussed and studied later in this chapter. For Mill Creek the MILUB farm income difference (DNRl = $258,042), 35% is achieved at time zero (beginning of year 1) with instant inStallation, another 25% is added in equal increments (5% per year) over years 2-6, and still another 20% is added in equal increments over years 7-16 (2% per year), with flows counted as occurring at the beginning of each year. The sum of these percentages is 80% which is the maximum P1,if,t’ and the farm income difference (DNR) reaches the level of $206,434 ($258,042 x .80) in years l6-50 of the evaluation period. The other components of EB grow in similar fashion. Alternative Investment Criteria As previously indicated, the SCS B/C is a computational routine that provides numerical benefit cost ratios that are the equivalent of those obtained using SCS procedures, as in chapter 3. Other kinds of invest- ment criteria used in this chapter include the IRR (internal rate of return), sum of NPV's (net present values) for all 12 projects, and the PV B/C (present value benefit cost ratio). One difference between the SCS and these other criteria is that SCS employs amortized capital cost flows, whereas the other criteria employ capital cost flows. Another difference is that SCS uses multiple interest rates in each project *r ‘fi w 3USDA, SCS, Watershed Protection Handbook (Washington, D.C.: SCS, 1967), sec. 102.0211. 220 evaluation. This is done only with the SCS NPV among the alternative, non—SCS criteria, as will be described later. Single interest rates are used in other criteria that are alternative to the SCS criterion, but it should be pointed out that multiple rates may be entirely appropriate from a policy standpoint; they can be used with NPV and BIG computations, and even with Robert Marty's modification of the IRR, called the composite internal rate of return, as described briefly in chapter 4.4 The question of how to develop suitable capital cost flows is more diffi- cult, and this will be taken up following the general introduction of the alternative, non-SCS criteria formulations. Following McKean and others, the author will define investment to mean a negative net cash flow (bt - Ct)’ A positive net cash flow is counted as a receipt. In this definition there is no distinction between capital costs (SCK and ACK) and recurring costs (Operations maintenance costs, SCOM and ACOM). Capital cost (SCK + ACK) is not 5 the same in meaning as investment. Thus, (net cash flow)t = (b — Ct)’ without reference to whether t Ct represents a capital cost or recurring cost or both. More formally: 3 (net cash flow)t = AN + if [(DNR - FWDCif - ACOMif) X Pl’if’t if ‘ (ACK1£,1 + “11133) X (P1,if,t " Pl,if,t-l)] 4On the use of multiple rates to distinguish publicly and privately borne benefits and costs, see KennethrJ. Arrow and Robert C. Lind, "Uncertainty and the Evaluation of Public Investment Decisions," American Economic Review vol. 60, no. 3, June 1970, pp. 364-387. Also, see Robert Marty,“The Composite Internal Rate of Return," Forest Science, vol. 16, no. 3, September 1970, pp. 277- 279. SRoland McKean, Efficiency in Government Through.8ystems Analysis (New York: John Wiley and Sons, Inc., 1958), pp. 76-77, 114-116 and 122. «I: H.....K....»Pl abet 221 If t > 1, AN = FWDRB + OTHERB - SCOM. If t 1, AN = FWDRB + OTHERB - SCOM — SCK. If t = l, P1,if,t-1 = 0. An exception occurs if the SCS instant-installation assumption is dropped, in which case an SCKt is specified for years 1-5. If (net cash flow)t is negative, it is added to the C (investment) present value sum; if it is positive, it is added to the B (net receipt) present value sum. T _ B a 2 [(positive net cash flow)t / (1 + r)t 1] C II ”Mt-1 ('1' [(negative net cash flow)t / (l + r)t-l] For any one year (t) the net cash flow will be either positive or negative. Thus, each year will contribute a net cash flow to either the B (net receipt) sum or the C (investment) sum, but not to both. This is emphasized because other definitions are possible. Using B (net receipt sum) and C (investment sum) as defined here, the present value benefit cost ratio PV B/C = B/C which is not the same as SCS B/C. The net present value NPV = B - C. The internal rate of return IRR is the rate of discount for which NPV 8 0!; As previously mentioned, the author computed data for the SCS NPV which is a net present value, like the NPV just discussed, except that SCS-assumed multiple interest rates were used in place of a single 6The computed IRR's are approximate and are associated with a slightly positive, rather than zero NPV. They are the lowest discount rates with a positive NPV computed to the nearest 0.1 of a perCentage point. For example, if the computed IRR a 5.0%, the NPV is slightly positive, and for r = 4.9%, the NPV is negative. 222 interest rate. Interest rates r1, r2, r3 and r4 are employed in a manner that appears to be compatible with SCS usage, although complete compatibility is not possible. Rates r1 and r2 are used to discount ACKl and ACKZ capital cost flows respectively. Rate r3 is used to discount non-ACK components of EB (DNR, ACOM and FWDC). Rate r4 is used to discount FWDRB, OTHERB and SCOM, all of which are constant for all years in the evaluation period. Complete compatibility with SCS assumptions is not possible, because SCS amortizes ACK's using rates r1 and r2, combines these amortized ACK cash flows with other EB annual cash flows (DNR, ACOM and FWDC), discounts the combination at rate r3 for summation into a present value, and then amortizes this EB present value into an annual equivalent value using rate r3. Furthermore, there may be some question as to whether rate r3 or rate r4 should be used to discount FWDRB, OTHERB and SCOMB to best reflect SCS assumptions, although this question does not arise for the more recently evaluated of the 12 studied projects, since for these projects r3 3 1'4. ACK Investment Rates for Alternative (Non-SCS) Investment Criteria SCS applies the analogs of EB achievement rates Pl,if,t to the combined EB annual cash flows (EB—100% in Table 3.8, chapter 3, i.e., the alebraic sum of DNR, ACOM, FWDC and amortized ACK) in a way that can be readily used in formulating SCS B/C, as described earlier in this section. In effect these combined EB cash flows increase after the fashion of a decreasing—rate growth curve during the EB development period, after which they become constant at a rate determined by the maximum value of the EB achievement rate variable Pl if t (see 9 9 223 Figure 7.3). For the alternative, non—SCS investment criteria, non-ACK or recurring annual cash flows (DNR, FWDC and ACOM) follow the same pattern during the evaluation period. However, the ACK cash.flows follow an investment rate pattern based on the EB achievement rate difference Pl,if,t — P1,if,t-l (with Pl,if,t-1 the EB development period, after which they become constant (at a = 0 for t = 1) during zero or other rate, as explained later in this chapter). Thus, for the alternative, non-SCS investment criteria the ACK investment-rate cash flows decline gradually during the EB development period. Disregarding non-ACK cash flow components of EB, undiscounted ACK cash flows for any year t for the SCS B/C are obtained as: ACKt = if {[(ACKif’l x A1) + (Acxif’z x A2)] x P1,1f,t} For the non-SCS investment criteria (P1,if,t—1 = 0 for t = l): 3 ACK = 2 CK + CK - t if [(A if.1 A if,2) x (Pl,if,t Pl,if,t- the ACK economic lives equal the evaluation period in years (as 1)] assuming discussed later in this chapter). In the SCS B/C the ACKt is an amortized annual cash flow, reduced by the proportion Pl,if,t9 and in the non-SCS investment criteria the ACKt is a portion of the total capital cost investment (ACK, ignoring SCK). For example, suppose: ACK 8 $1,000,000, one EB component (thereby summation over subscript "if" is superfluous), and amortization at 6% interest over 50 years (amortized ACK cash flows, ACK x A = $63,440). And Pl,if,t’ .35 at time zero (first part of year 1), increases .25 over years 2'6 (.05 per year beginning one year hence), and increases another .20 over years 7vl6 (.02 per year), for a maximum Pl,if,t of .80 for years 16-50 of the evaluation period. ACK cash flows are as follows: 224 Year P . ACK for SCS B/C, ACK for non— SCS criteria - 1.1f.t t n. ....... (amortiZed) investment rate) 1 .35 $22, 204 $350, 000 2 .40 25,376 50,000 7 .60 39,333 20,000 16-50 .80 50,752 -0- Amortization, Cash Flows and Discount Factors For both the SCS and alternative, non-SCS investment criteria, the present values of discounted ACK cash flows are obtained as follows: PV of ACK cash flows = % [ACKt / (1 + r)t-1] At 6% interest the sums are $702,981 (using amortization) and $670,616 (using investment rate ACKt data), reSpectively. Actually, either could be used in computing an SCS annual benefit cost ratio, but the equiva- lent of the former is used by SCS, and it results in a lower ratio than if the sum $670,616 is used, disregarding other cash flows. The difference has to do with the technical point that the discount-factor power t is more consistent with the definition of amortization than the power t-l. However, any overstatement of ACK deductions (reducing the benefit cost ratio) is outweighed by the effect of using the discount- factor power t-l rather than t on all EB annual cash flows for years t = 1-T (where T = 50 or 100), thereby increasing the benefit cost ratio. Recall that SCS used as many as four different amortization factors in evaluating some of the 12 studied PL 566 projects. Two of these, A3 and A4, appear to be rather straightforward in effect, since they both are used to convert a present value into an annual equivalent for use in the SCS annual benefit cost ratio. However, Al and A2, even A4, diStract attention from critical variables, the capital cost investment 225 rates. The numerical data given in the preceding paragraph compare the present value of ACK amortized cash flows ($702,981) and ACK investment rate flows ($670,616), but both flows are based on the SCS enhancement benefits achievement rates Pl,if,t' Do enhancement benefits achivement rates Pl,if,t represent a realistic specification of ACK cash flow rates, that is, ACK investment rates? SCS work plans specify a much faster rate and more complete degree of land treatment investment than do the Pl,if,t data for ACK. The difference is significant, as will be shown later in this chapter.7 Technically, using the discount-factor power t rather than t—l is consistent with the definition of amortization. For example, given an amortized flow (at interest rate r, for N years) of $20, this can be obtained as $20 = Ar,N x g[$20 / (l + r)t], whereas the discount-factor power t-l would give a value larger than $20 on the left side of the equation. The amortization factor is computed as follows: Ar,N = r [1-1/(1+r)N] 7The ACKt cash flows for any year t are based on the EB achievement rates P1,if,t for the SCS BIG and on the achievement rate difference P1 if’t - Pl’if,t-1 (with Pl’if,t—1 = 0 for t = l) for alternative, non- SC$ criteria, as explained earlier in this chapter. SCS procedures, as incorporated in SCS B/C in this chapter, apply EB achievement rates Pl,if,t to amortized ACK cash flows which are a part of the entity of all EB cash flows (DNR, ACOM, FWDC and amortized ACK cash flows). This may be done as a matter of simplification by SCS. When transferred to non- SCS investment criteria the EB achievement rate differences (Pl,1f,t - Pl,if’t_1, with Pl’if,t-1 = 0 for t = 1) provide a rather slow and par- tial completion of ACK investments. Much faster rates of completion (5 year as Opposed to 15-30 year) and higher degrees of completion (100% as opposed to about 80% typically) are shown in typical SCS work plans for the land treatment investment in a watershed. These land treatment investment rates are applied to ACK investments in the section of this chapter on cash flow timing and instant installation. Project NPV's, IRR's and B/C's are reduced significantly. 226 For comparison, using the power t in forming the present value sum of ACKt cash flows in the preceding example gives the sum $663,190 and this is less than the sum using the discount-factor power t-l, $702,981, where ACKt is an amortized flow (ACK x A6%,50 yrs. x Pl,if,t)‘ The lower of these two sums is reasonably close to the sum $670,616, which is obtained using the power t-l and investment-rate ACKt cash flows [ACKt = ACK x (Pl,if,t - Pl,if,t-l)2 where Pl’if,t_1 = 0 for t = 1]. However, for consistency, the author used the discount factor t-l in both SCS B/C and the alternative, non-SCS investment criteria. The discount-factor power t-l seems to reflect SCS procedures as used in the 12 studied projects, as discussed for the Mill Creek project in chapter 3 (see Table 3.8 where 35% of $120,211, the EB-100% value, is counted at full value, the equivalent of counting a cash flow at time zero, the beginning of year 1, meaning the discounting of cash flows with the discount-factor power t-l and not t). Again, another point to keep in mind is that using the discount- factor power t-l increases the present value of any net cash flow (b - C)t' Given the typical cash flow patterns for a PL 566 project, this results in a higher numerical value for the SCS annual benefit cost ratio or any other investment criterion than if the discount—factor power t is used. Some investment analyses use the discount-factor power t-l for capital cost cash flows and the power t for project-credited income. The rationale is that capital cost flows occur at the beginning of the year and income later in the year. Implicitly, this means a causal relationship between the two flows for one year. As indicated in chapter 3 and later in this chapter, capital cost cash flows are in 227 themselves insufficient to result in the projectacredited farm income, because management choices involving changes in crop input combinations, use rates and practice timing are also necessary. Furthermore, SCS assumes changes in output combinations. One might expect some increase in income due solely to the capital investments, but not as much as SCS credits to a project, based on the assumption that these several complementary changes will occur. To summarize, SCS uses amortized ACK data and computes an annual benefit cost ratio in such a way that obtaining present value sums by using the discount—factor power t-l rather than t is implied. While this decreases the benefit cost ratio in terms of the effect of ACK alone and is inconsistent with the definition of amortization, the overall impact is to increase the ratio, since net cash flows for any one year are generally positive. Thus, cash flows for all years in the evaluation period are counted as occurring at one year intervals beginning at time zero (the beginning of year 1), then one year hence (the beginning of year 2), then two years hence (at the beginning of year 3), ..., and finally T-l years hence (at the beginning of year T). Some economists recommend discounting capital cost flows by the power t-l and other annual cash flows by the power t, assuming within-year causality. Any one of these differences will affect the apparent worth of a project to some extent, regardless of the investment criteria used. In the author's judgment one of the most significant problems with.the use Of amortized ACK.cash.flows is that the combined EB cash flows (the algebraic sum of DNR, FWDC, ACOM and amortized ACK) are treated as a single entity with a single accrual or growth pattern, whereas there is reason to believe ACK investment cash flows are made 228 at a much more rapid rate than is suggested by the EB achievement rate data Pl,if,t' This is not to deny the relevance or impact of any of these other possible changes. Post EB Development Period ACK Investment Rates for Non-SCS Criteria As previously indicated, ACK cash flow or investment rates for alternative, non-SCS investment criteria are determined by EB achievement rates for any year t as follows (with Pl,if,t-l = 0 for t = 1): 3 ACK: = EfHACKfia + ACKif,2) x (Pl,if,t " Pl,if,t-l)] Recall that the EB achievement rate Pl,if,t reaches a maximum after the EB development period, meaning after the first 15-50 years of the 50-100 year evaluation period, depending on the project. In the previously cited example, the maximum Pl,if,t = .80 occurs in year 16 and remains constant over years 16-50. For these years no ACK investment occurs since the ACK economic lives equal the evaluation period in years. However, if the ACK investment has an economic life shorter in length than the evaluation period, replacement investment is required. In the SCS B/C this is handled by using an ACK amortization factor based on a shorter period of time. For simplicity the author assumed for non-SCS criteria a constant rate of ACK replacement investment flows for projects in which SCS assumed that the ACK economic life was less than the evaluation period in years. Other patterns of replacement investment could have been assumed. Essentially, the eventually completed amount of ACK investment is re-invested after the initial ACK life has been completed. 229 Continuing the previous example, suppose that the ACK economic life were 25 years instead of 50 years, then ACKt=26-50 = $32,000. That is, 80% of the initially estimated ACK investment of $1,000,000 is assumed to be completed eventually, and this along with the ACK economic life determines the amount of the replacement ACK investment for years 26-50 as follows: $32,000 = $1,000,000 x .80 / 25 year ACK economic life More generally, ACK replacement investment rates for years beyond the initial ACK economic life period are determined as follows: ACKt = §f[(ACKif,1 + ACKif,2) x (maximum Pl,if,t) / (ACK life)] for t>(ACK economic life) to t = T. This explanation has been simplified a bit in that the on-farm (ACKif,1) and inter—farm (ACKif,2) investments require separate treat- ment in some cases. Actually this computation is required for only 3 of the 12 studied projects to show the effect of SCS assumptions in the base estimates for the non-SCS investment criteria (Table 7.1). However, the computation is required for all 12 projects for certain sensitivity analysis alterations of SCS assumptions, as discussed later in this chapter. Maximum or eventually-completed EB achievement rates Pl,if,t are shown in Appendix, Table 2 (columns 12 and 16), along with the SCS- assumed ACK economic lives (columns 5 and 7) and evaluation period lengths (column 3). .figse—Estimate Investment Criteria Data Computed data for several investment criteria are shown in Table 7.1. The SCS B/C computational routine provides annual benefit 230 was mmumu vmaswmmlmom one home mm nomads has moumu umoumuoa bosommmlmum manna .N wanwa .vaomom< ca ozonm .muomfioum macaw ummmfiw mumm mum» osu use .uomhouo Hon usom mm .moumu unwound“ mHoHuasa uoaomwm Imom msu moans wousmaoo one some >mz mom mnu was Amoaumu umoo ufimmomn Hanson vmusmaoo -mum was mumaaxouaam assess mums o\m mom was soon .uxmu name was as emaamaaxm was .MIH moaan .xaeaoom< .mumu usaofi amuwoun .vmuoaaoo "mousom oae.aow som.mom masses Hmm «mm Nq.m oh.H -.N memo onus Hmo.a Ham Nm.m m~.~ oo.~ seas commuaum mon.m qsm.q N~.~N H~.m mm.s Nomfl amuse Hafiz ass mom Nm.~H ss.~ sa.¢ ammo amusmaz Hoo.m HsN.o Nm.sa a~.m No.0 ammo smswmumflz mm o NH.m mo.H mm.a mesa mousse cam so N~.o «N.H Hm.H soma samba om am N- Nw.q as.o -.H mesa sumac sums ans mm Ns.o sm.a ma.o coma assume mmq.H mmo N~.o o~.H mo.H Homo .m .mmmo o~o.a mma NH.a so.~ NH.m ness .2 .mmmo ANN a ems m um.a wo.~ mm.~ memo somam m . 000.; m . OOOHm mMH OHumH NOHumH mumv mama HUMHOHA m>mz mom an on .Nm um .o\m mom uoofioum >mz o\m >m .wouamao .muumfioum son as ammaauaz «H .mqu unassumm «mam «sausage uaoaumm>aH .H.~ «Home 231 cost ratios that approximate those computed using SCS procedures, although the two routines differ as explained previously in this section. The SCS B/C ratios of Table 7.1 and those computed by SCS are shown in Appendix, Table 3, and there is some discussion of the relatively minor differences between them in chapter 1. The PV BIG and the NPV data shown in Table 7.1 are computed at 5%.8 They are computationally con- sistent with the IRR, meaning that the PV B/C = 1 and the NPV = 0 when computed at an interest rate equal to the IRR. The SCS NPV is an NPV, like the NPV just discussed, except that the SCS-assumed interest rates are employed (with as many as four per project and with interest rate sets differing among projects). Ranking Projects The 12 Michigan PL 566 projects are ranked in Table 7.2 using different investment criteria as ranking devices. However, the results should be interpreted with caution. Recall that SCS evaluated these 12 projects over the period 1959-68. Crop input and output prices and coefficients differ. Growth patterns for enhancement benefits differ. Capital costs reflect different price levels. The problem of different interest rates is removed for the non-SCS criteria, except the SCS NPV which uses the original SCS-assumed rates. SCS-assumed ACK economic lives differ. With these differences in mind the project rankings may be considered to illustrate the general effect of ranking by various criteria. 8The NPV and PV B/C were obtained for 20 interest rates, but the NPV and PV B/C data are shown here only for a rate of 5%. The 20 rates are as follows: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, l4, l6, 18, 20, 25, 30, 35, and 40 percent. 232 .m.h magma ou wsfiumamu one» Gama.QAu ow mm Ham: on .olm mmuosuoow .m.n manwa as haasu whoa vooflmamxm «para a o\m mom owuwu onu mwfi>oum wasos .moumu maowuaoa woasmwmlmum no woman :H com: «H enowms .mouwu amoumusw meon vow moqumu Onma o\m momn .q manna .Nfivoomo< ca ameaw mum maoaumooa wow moamo Hash .oan .0969 wow ”communum .uoum mxoouo Hafiz .HHHS mumuxmsz_.xmez mwmswouwfiz .umaz moHuuHA .uuHA mow .ow «sumac spam .summ mosaumo .Humo m.m .mmmo .mmwo «.2..mmmu .Zwmo “Momam .omam “oonHom mm mum moans uomfioumm .MIH moanma .Nauomma4 aw pump mamas summons .vmuomaoo “mousom 7 F r [Ll mm.o sums mm.o been N~.H some m.q sums m.s sums N- sums NH mo.H “has as.o ow mm.H duos H.m “has H.m boss 0 boss Ho NN.H on am.o Homo Hm.a ow H.s ow N.o ow am Home as o~.H mmao mm.o anus so.a mmmo H.o mmmo ~.c mama so ow a sm.H Homo oo.H mmmo mn.o Homo m.o Hume “.0 Homo ems umam w on.H onus so.a onus co.~ “sum o.m “sum s.w onus mom amp: 5 mo.~ ammo mo.H “sum N~.~ came m.m came H.m ammo Ham “sum 0 wo.~ omam om.H ammo mm.~ umom s.w ammo m.m snow «mm onus m mm.~ “sum am.a umom NH.m same ~.m umom m.s swam mmo mmmo s sq.~ sass No.m smaz mm.q awn: m.HH ems: m.~H own: man ammo m s~.m “we: om.s “we: mm.s Hafiz. m.~H owe: m.sH Baez same Hafiz N HN.m Hafiz sa.s Hafiz na.m owes m.mo Hos: s.- Hoax Heme “was H oaumu neous“ . oaumo N N m.oooom seam Nm um ammo . , ommn no u o\m mom mmH an o\m >m o\m mum o\m mom ,uom .u mowaam um >mz I n.mmauouauo usoaumo>oH venomaom we umxsmm muoonoum oom.qm amenofiz meansh .N.m magma 233 The explanation for inconsistent project rankings between the IRR and NPV's computed at various interest rates may be found in Figures 7.1 and 7.2; that is, NPV curves differ in responsiveness to changes in interest rate and they may intersect before reaching the IRR (the interest rate at which NPV = 0). Similarly, the PV B/C curves for different projects differ in reSponsiveness to changes in interest rate and may intersect before reaching the IRR (the interest rate at which PV B/C = 1). Thus, one should not expect NPV and IRR rankings and NPV and PV B/C rankings to agree. Furthermore, part of the divergence between NPV and PV B/C rankings may be explained by the tendency of PV B/C curves and NPV curves to interesect respectively. However, the PV B/C is a ratio (present value of benefits / present value of costs), and NPV is a difference (present value of benefits - present value of costs). On this basis, one would expect differences in project ranking between the NPV and PV B/C. Despite some criticisms, the IRR is a useful device. It is an analog to the single interest rate, which, if used in place of the SCS- assumed multiple rates of interest, would make the SCS BIG 8 1. Their relationship will be discussed in detail later. Only 3 of the 12 projects change ranking between these two devices (Table 7.2). The NPV and PV B/C data in Table 7.2 both assume an interest rate of 5%, but ranking divergences are clearly apparent. On the other hand, the base SCS B/C and PV B/C (at 5%) data provide somewhat similar rankings. Project rankings with the SCS B/C 1970 data differ from those with the SCS B/C base-estimate data. The SCS B/C 1970 data will be discussed in more detail shortly; briefly, they reflect interest rate and ACK economic life assumptions that might be used if the projects were re-evaluated 234 by SCS in fiscal 1970, other things remaining unchanged. Projects change rank rarely by more than one or two places in Table 7.2 until the NPV ranking is compared to the ranking by any other device (excluding the SCS B/C 1970 which introduces a change in underlying variables). Interest Rates Background The question of what interest rate (or rates) to use in evaluating public investments has received considerable attention by economists, but there are several unresolved issues, as discussed in chapter 4. Social opportunity cost proponents argue for a rate (or rates) usually representing the return on foregone uses of the real resources. De- pending on the means of measurement, rates ranging from perhaps 7% to over 15% might be suggested for 1969-70. At the other end of the scale, a social time preference rate, roughly represented by the yield on long-term Federal bonds (6.8% for 1969), with a reduction to remove the market inflation-expectations, could be in the vicinity of 4-5%. Owing to the problems of market imperfections and uncertainty and a host of other difficulties, it does not appear that the two rate positions can be reconciled. In the selection of a rate three general issues stand out: capital formation (or net investment, having to do with the rate of economic growth), the division of investment between public and private projects, and the division of resources between short and long term projects. To encourage a higher rate of capital formation, the government could lower all interest rates via monetary, fiscal and debt-management policies. 235 Capital formation may be accomplished either in the private or public sectors, but to transfer resources from higher to lower earning potentials is inefficient. Although there may be some questions about definitions, the internal rate of return (IRR) on the projects studied in this chapter may serve as a rough yardstick for comparison with earnings in the private sector. Individually, only three of the twelve projects have an IRR above 10%. Taken as an aggregate, the twelve projects have an IRR just under 12%. Rates of return on private invest- ment may be in the 10-20% range, depending on definition and the risk class (variability of outcome class). Could not the twelve Michigan PL 566 projects have been good investments, allowing for the rise in rates of return since the early 1960's, when most of these projects were planned and approved, and allowing for the lower rates of return on utility-type investments (due to less variability of outcome)? Perhaps, but the project rates of return stated here are base-estimate rates (as shown in Table 7.1) and no corrections have been made for optimism bias. Possible sources of agency optimism bias in stating project returns have been studied in chapters 5 and 6; they will be taken up again later in this chapter. Depending on the assumptions selected to judge the agency estimates, this optimism bias may be slight or considerable. In any event, if returns on the public projects are lower than returns for comparable private sector investments, a misallocation of resources has occurred, and society would have been better off to avoid such public projects. Concerning the division of investment between short and long term projects, one position is that private market rates of return on investment are high because individual (private) decision makers are 236 short-sighted. It is presumed that society should react differently, paying more attention to the future to offset the defective telescopic faculty of individual decision makers. In other words, the government could use a lower rate of interest in evaluating public investments, thereby emphasizing long-term projects. Lower interest rates have the effect of increasing the relative value of benefits farther in the future than do higher rates. Another way of saying this is that lower rates encourage capital-intensive, long-lived investments. Assuming that the distant-future benefits are in fact "correctly stated" (i.e., assuming no optimism bias), critics of low discount rates argue that an income transfer from the poor to the rich is implied, since real per capita income increases through time. Furthermore, the use of low discount rates is not a means of encouraging public as opposed to private investment, if this is desired to counteract the short- sightedness of private decision makers on the aggregate rate of capital formation. Rather, low discount rates encourage capital-intensive public investments; that is, for example, spending on school buildings rather than education. (Note: capital intensity and economic life may refer to different dimensions of PL 566 project investments, as discussed in chapter 8.) AgencyiPractice As mentioned earlier in this chapter, SCS has evaluated PL 566 project investments using as many as four different interest rates per project, with the rate sets changing with projects through time. Since about 1965 the SCS Planning Party in Michigan has used just two rates 237 per project.9 The inter-agency authored GreenbOok (of 1950) apparently accounts for the two rates used on the earliest planned of the 12 studied Michigan projects (the Misteguay, 1959). For the Misteguay, the SCKramortization factor (A4) was based on an interest rate (r4) of 2.5%. In contrast to the 2.5% rate for SCK amortization, the Greenbook advo- cated a 4% rate for handling private costs and benefits (meaning r1 = r2 - r3 = 4%), as for the Misteguay project. Eckstein is critical of the Greenbook's recommendations of a 4% rate, if the reason is the evaluation of "private" costs (meaning associated costs, ACK, which may be public costs in part via Federal ACP payments to farmers). However, if the reason is to distinguish private and public risks, Eckstein is willing to accept use of dual rates. More recently Arrow and Lind advocated the explicit use of dual rates, a lower rate for evaluating publicly borne costs and benefits, and a higher rate for evaluating privately borne costs and benefits.10 Since 1962 evaluations have been based on the rather ambiguous policy statement known as Senate Document 97, actually authored by an inter-agency committee and later published by the Senate. It recommends 9As shown in Appendix, Table 2, four of twelve projects were actually evaluated with r3 # r4, although the work plans indicated r3 = r4. This may mean that enhancement benefits were computed when the rates for one fiscal year were in force, and that SCK's were amortized later when higher rates for the next fiscal year were in force. 10The Greenbook: U.S. Congress, Subcommittee on Benefits and Costs of the Federal Inter—Agency River Basin Committee, Proposed PraCtices for Economic Analysis of River Basin Projects’... (Washington, D.C.: USGPO, 1950). Otto Eckstein, Water Resource Development (Cambridge, Massachusetts: Harvard University Press, 1958), pp. 60, 64, and 88-90. Kenneth J. Arrow and Robert C. Lind, "Uncertainty and the Evaluation of Public Investment Decisions," American EcOnomiC‘Review, vol. 60, no. 3, June 1970, pp. 364-378. 238 a single rate for evaluations. However, SCS practice (1967—70) is based on two rates, with a second, higher rate for ACK.amortization, following the concept of local borrowing costs.11 Economists have been critical of Senate Document 97's definition of a long-term rate for use in evaluations. Essentially, it is a coupon rate for outstanding, marketable Treasury obligations, which were long- term at the time of original issue, but which may be due or callable in a much shorter period of time. In 1968 the agencies redefined the rate to relate to current yields on such obligations with maturity or call dates 15 or more years in the future. Thus, the definitions differ on two points. However, the new definition rate was established at a base of 4.625%, below the then current yield on long—term bonds, and it is not allowed to advance more than 0.25 of a percentage point per fiscal year. Thus, a rate of 4.875% was in use in fiscal 1970 (July 1, 1969 to June 30, 1970) when the current yield on the bonds was closer to 7%. Apart from the criticism of the water resources agencies' definition of a long-term interest rate, economists who advocate "risky" or opportunity cost rates may find the rate too low for other reasons. This is because yields on long-term government bonds (with a deduction for inflationary expectations) represent an essentially riskless, lenders' rate, which is perhaps a reasonable approximation of a social time preference rate. w ‘w—w‘fiVa‘fi- V 11U.S. Congress, Policies,'Standards and PrOCedures‘..., Senate Document 97, 87th Congress, 2d Session (Washington, DiC.:11USGPO, 1962), pp. 11-12.‘ USDA, SCS, Economics Guide (WaShington, D.C.: SCS, 1964), ch. 6, p. 5 and ch. 7, pp. 6—7..-On SCS interest procedures, see USDA, SCS, Watershed Protection Handbook (Washington, D.C.: SCS, 1967), sec. 102.0213. 239 Single—Rate Proxies for SCS Multiple Interest Rates Table 7.3 shows several discount rates for the 12 studied PL 566 projects, both single-rate proxies for SCS-assumed multiple interest rates and internal rates of return (IRR). Each is computed under two different sets of assumptions, those used by SCS in the original evaluations and those which might have been used by SCS in fiscal 1970, other things remaining unchanged. For 1970, 25-year ACK economic life is assumed, while the original SCS evaluations used ACK economic lives ranging from 15 to 50 years. SCK are amortized using an interest rate of 4.875%, and ACK are amortized using a rate of 6.5%. These interest rate assumptions do not affect the IRR's and single rates (to make SCS B/C = 1), because the IRR's and these single rates are program outputs and not data inputs. When the single rate proxy for SCS multiple rates is determined for a project, the computations are iterative, employing one rate at a time until a rate is found to provide the required SCS B/C ratio. That is, the procedure described in the methodology section of this chapter is used to compute the SCS B/C, except that r1 = r2 = r3 = r4, with various rates being tried until the one is found that provides approximately the correct SCS B/C (as shown in columns 2 and 3 of Table 7.3). The same process is used to obtain the rate for which SCS B/C = 1. Computation of the IRR is also iterative, but the procedure finds the rate for which the NPV = 0 and the PV B/C = l. The 1970 interest rate and ACK economic life assumptions signifi- cantly reduced the benefit cost ratios for all 12 projects. Similarly, the single—interest-rate proxies for the SCS multiple rates rose 240 .mummm on ou ma scum muomhouo wsoam moaum> muss ofiaoaoom M04 Amum Hosawauov opossumm omen ecu newness .mumm% nu ma muomhoum Ham now muss Mu< Ohms was monsoon Mommaw As was w woaoaoov m.mmH com Am use 0 moansoov H u o\m mom oxwa ou muuou mamaamo .u\m mom uamHm>Hsvm wnu weseoum vanes .mmumu mamfiuasa Amum umHHmaH omma uov umBSmmmlmum msu mo woman ow now: «a .noHnB .ummuouGH mo «mums mum meNoum mums ummumuoa mamasmn .uomhosa an “amuse Amolmmms posses emu Homv moosueaowmm “mum HmusHHOV ouossumm omen msu newness .muoonoum saw you wamm ago was msosumssmmm owonH .mummm nu ma muss usaosoom M04 vow .Nm.o mum Ann was an mmumuv oosumususosm Mom How momma ammumuos .Nmmw.e mum Aasm>auooammu .nu use an mouwuv msowumusoaoo mm use sosumusuuoam Mum you money unassuss ”msossom on use moosumsdmmm ohms ona .mBMm onu wasosmsou mwssnu nonuo .onms Hmomsm ow woumosw>otmu one: muomhosm mam ms .oasmmm uana mow umgz umowwsm ou venomous mum mooHuasbmmm cums one .N mamas .Nsvsomm< as ssoam mum msosueasmmm Amom Hmsswsuov opossumo ommmm .mts woanma .vaammew .musmss muse amuwoum .vouoaaoo “mouoom a e e e s e a.e m.e a.a ee.s NN.N eeas ones e.a a.a a.e o.e s.e e.a ae.s oo.~ eeas commusea e.~a a.- a.as e.as s.e o.a aa.e ea.a seas Mambo sssz e.ss a.~s e.os a.ss a.e a.a ~o.a aa.e aeas nausea: a.ms m.es a.ss m.ss e.e o.a oe.e as.a aeas amsmmumsz ~.e s.e e.e s.e e.e s.m ae.o ma.s Neas assess o.e N.e e.e s.e e.e s.e sa.o se.s eeas usage as a.e e.e o.e e.e s.e e.a ea.o Na.s eeas samba apes e.e s.e e.e e.e a.e e.e sa.c es.s eeas assume e.e N.e e.e s.e e.e e.m oo.s ae.s seas .a .aamo s.e s.a a.a a.e e.e s.e oe.s as.m aeas .2 .memo Na.e Ne.a Ne.e s~.a Nm.e Na.m se.s ee.~ aeas sense a e a e e e - a - N s OmmH mama OmmH mmmn omma mmmfl OfiUMH OHumH mummy GOHumquafiuow unknown «0 mums HmosmucH oa,n u\m_mum.mmm o>m mom now Gama omen paw mama uomhoum ‘7’}? P ammMNosm Mums “moumuss msmssm.L .o\m mom o\m mum r yr if n.mmuomnoum com Am ommsnoaz NH .auoumm mo mmumm HmsnmuaH use mmfixoum mama ummuoudH mamcsm .m.m mHQMH 241 (compare the rates in columns 4 and 5, Table 7.3). Single-interest—rates (to make SCS B/C = l) and IRR's fell for 10 of the 12 projects because the assumed ACK economic life (25 years) is lower for the 1970 computa- tions. However, these two rates do increase for the Farm Creek and Sturgeon projects, because SCS assumed ACK economic lives shorter than 25 years (see Appendix, Table 2). Why do the IRR's and the single interest rates (which, if used in place of the SCS-assumed multiple rates would provide an SCS B/C = l) differ? First, slight computational discrepancies could account for differences of perhaps 0.1 to 0.3 of a percentage point. Second, IRR's are generally higher than the single interest rates (to make SCS B/C = 1). An explanation has to do with the use of amortized ACK cash flows rather than investment—rate ACK cash flows in the SCS B/C. Recall that in the methodology section it was shown that the sum of amortized ACK cash flows is larger than the sum of investment rate ACK cash flows (assuming r = 6%, EB achievement rates Pl,if,t for MILUB for the Mill Creek project and ACK = $1,000,000, the discounted present value sums are $702,981 and $670,616, respectively). This difference would lower the computed SCS B/C at any rate of interest as compared to what the SCS B/C would be if investment rate ACK flows were used as in the non-SCS investment criteria. Thus, in general, if the SCS B/C and PV B/C were computa- tionally consistent, one would expect the SCS B/C to reach a computed value of one at a lower rate of interest than the PV B/C. Third, contrary to this explanation, the single interest rate to make SCS B/C = 1 exceeds the IRR for 4 of the 12 projects under the 1970 assumptions specified in Table 7.3. Apart from computational discrepancies some other factor could account for this. Fourth, the SCS B/C has a 242 different cash flow arrangement than the nonascs investment criteria. One should not necessarily expect the SCS BIG and the PV B/C to provide a ratio of unity at the same interest rate. Interest Rate Sensitivity_Ana1ysis Net present value (NPV) curves for the 12 studied PL 566 projects are shown in Figures 7.1 and 7.2. In Figure 7.1 four rates of return over cost (RRC's) may be observed, that is, discount rates at which the NPV curves for two projects intersect (see chapter 4). If NPV curves did not intersect, IRR's (discount rates for which NPV = 0) would be valid ranking devices, regardless of the interest rate selected. For example, at discount rates above about 9.5%, the NPV for the Mill Creek project exceeds the NPV for the Misteguay, but for discount rates below 9.5% the Misteguay has a higher NPV. While the NPV variations among the 12 projects are too great to conveniently draw all 12 NPV curves on one set of axes, one can still observe different degrees of interest rate responsiveness among the NPV curves. A11 NPV curves are more reSponsive at lower rates of interest (that is, the curves are steeper). However, notice that the NPV curve for the South Branch of Cass River project does not level out quite so well and forms RRC's with the NPV curves for six other projects before reaching its IRR (6.4%, where NPV = 0), as shown in Figure 7.2. The NPV curves for the Sturgeon and Tebo projects intersect twice (RRC = 4% and RRC = 6.4%). Below the lower RRC (4%) and above the higher RRC (6.4%), the NPV and IRR rankings would agree, but not between them. Figure 7.1. Net present value (millions of dollars) Source: 243 NPV Curves, Michigan PL566 Projects. Mill Cs Mist Mill Mill Creek ” Mist Misteguay Cs Cass S. Cm Cass M. Musk Muskrat Cm p Musk l ‘— 5 10 Discount Rate (Percentage) Computed, program input data, Appendix, Tables 1-3. 244 Figure 7.2 NPV Curves, Michigan PL 566 Projects. T S Cm Cs Cs Cass S Cm Cass M S Sturgeon T Tebo M Muskrat 1.2-- J Jo Drain B Black C Catlin L Little M F Farm Creek l.O- A U) s w H '3 0.8 U J 1+4 0 U) \ s O 0H S 6 'E O. B v Q) s '3 c > 15’ 0.4‘ 33 L (D is D. i.) O) z; 0.2- F 0.0" 1— l I l 0 2 4 6 8 10 Discount Rate (Percentage) Source: Computed, program input data, Appendix, Tables 1-3. 245 NPV Sums for 12 Projects Because singleeproject IRR, NPV and BIG data require more space to present, lZ—project NPV sums at 5% will be used as primary quantitative measures of responsiveness to different changes throughout the rest of this chapter. However, the 12-project NPV sums may be used as discount- rate sensitivity indicators in their own right, as shown in Table 7.4, just as single-project NPV's can be used, as in Figures 7.1 and 7.2. For discount rates with positive NPV's for the 12 projects, notice that the NPV responsiveness to a given change, a 10% reduction in farm income (see Table 7.9), increases with the discount rate. Thus, the relative NPV data in Table 7.4 (column 5) suggests the importance of keeping in mind that NPV sensitivity at 5% is below NPV sensitivity at higher rates of discount, at least for this type of change. The higher lZ-project NPV-sum responsiveness at discount rates above 5% is due to the increasingly negative NPV's of some projects, as well as to the decreasing but still positive NPV's for other projects. Projects are not deleted from the sum if their NPV becomes negative. The use of a 12-project NPV sum at 5% may be defended as being based on a rate that approximates the marginal internal rate of return for the 12 projects. The marginal IRR is the rate of discount for which the lowest ranked project has a zero NPV. For these projects the marginal IRR is actually just over 4.8% (at which the Farm Creek project has a positive NPV of $115, compared to a $-1894 NPV at 5%). Table 7.4 shows the 12—project NPV sum using the SCS-assumed multiple rates of interest, as applied to their respective cash flow components. This is the sum of SCS NPV's which are mentioned in the methodology section of this chapter under non-SCS criteria. The sum of 246 $19.7 million approximates the single-rate sum at 4%, $18.2 million. Recall that the SCS-assumed interest rate sets differ by project. Table 7.4. Net Present Value Sums, Selected Discount Rates, 12 Michigan PL 566 Projects. NPV rela- NPV sum, for 10% farm Discount rate NPV sum, tive to income reduction (percentage) $ millions NPV at 5% $ millions Relativeb l 2 3 4 5 0 61.0 4.52 52.7 .86 1 44.4 3.29 38.1 .86 2 32.8 2.42 27.7 .85 3 24.4 1.81 20.3 .83 SCS-assumeda 19.7 1.46 16.2 .82 18.2 1.34 14.8 .81 13.5 1.00 10.6 .79 6 9.9 .73 7.5 .75 7 7.2 .53 5.0 .70 8 5.0 .37 3.1 .62 9 3.2 .24 1.5 .47 10 1.8 .13 0.2 .13 12 -0.4 -1.03 -1.7 4.29 14 -1.9 -1.14 -3.0 1.57 16 -3.0 -1.22 -4.0 1.32 18 -3.9 -1.29 -4.8 1.22 20 -4.6 -1.34 -5.3 1.17 25 -5.7 -1.42 -6.3 1.11 30 -6.4 -1.47 -6.9 1.08 35 -6.9 -1.51 -7.4 1.07 40 -7.2 -1.53 “7.6 1.06 Source: computed, program input data, Appendix, Tables 1-3. aSCS—assumed multiple rates differ by project, as shown in Appendix, Table 2. The NPV sum is the sum of individual project SCS NPV's as described in the methodology section of this chapter and shown in Table 7.1. bRatio = (NPV in column 4) / (NPV in column 2). 247 Enhancement Benefits Analysis Suppose that the SCS instant installation assumption is dropped. Then the EB (enhancement benefits) annual undiscounted cash flows may be visualized as growing after the fashion of a decreasing rate growth curve for the first few years of the evaluation period, after which they assume a constant annual cash flow rate, as shown in Figure 7.3. Actually, SCS assumed a three-segment linear approximation of the growth curve for several of the 12 studied PL 566 projects. As explained in the methodology section of this chapter, the SCS benefit cost ratio and the SCS B/C treat the algebraic sum of EB cash flows (DNR, ACOM, FWDC and amortized ACK cash flows) as an entity. This entity grows as shown in Figures 7.3 and 7.4 during the evaluation period. One significant difference between the SCS BIG and alternative, non-SCS investment criteria is the treatment of ACK cash flows. Recall from the methodology section of this chapter that the SCS B/C employs amortized ACK cash flows, with the flow for any year t based on the EB achievement rate P1,if,t' The alternative, non-SCS criteria employ ACK investment rate cash flows based on the EB achievement rate difference Pl,if,t - Pl,if,t—l' For both kinds of criteria the EB annual achieve- ment rate P1,if,t determines the cash flow for EB non-ACK recurring items (DNR, FWDC and ACOM). These effects of the EB achievement rate variable P should be kept in mind. l,if,t If the time axis were extended to include all years in the evaluation period (50 or 100 years, depending on the project), then the area under the EB growth curves in Figures 7.3 and 7.4 could be interpreted as representing the undiscounted present value of the EB 248 .N oHDmH .xsoswaa< mom on amDAHz .uomnopo xwwpo.sssz was now mumc no women Amumohv ease om os _ _ :Houuma nuzopw mm mums uomumsou \ noduwssxouaam unmewwmlm .pmossq ATIIIIII/III. w>uoo nuzosm mums wcsmmouooa om .mcuouumm nuzouu mm w>sumcumus< .m.n wuswsm OOH "mopsom 249 cash flows for SCS B/C (meaning DNR, FWDC, ACOM and amortized ACK). Similarly, the area under an EB growth curve can represent the undis- counted present value of some of the EB cash flows for alternative, non-SCS investment criteria, that is, only those directly determined for any year t by the EB achievement rate P1,if,t ( meaning EB cash flows DNR, ACOM and FWDC). Again, the ACK cash flows for any year t for the alternative, non-SCS investment criteria are determined by the EB achievement rate difference Pl,if,t - P1,if,t-l' That is, EB investment— rate flows ACKt decrease rather than increase during the EB development period. Figure 7.3 shows the EB growth pattern for the MILUB component (subscript if = l) of EB for the Mill Creek project, with the SCS instant installation assumption dropped. The achievement rate Pl,if,t reaches 35% of potential in year 5 (in equal annual increments of 7%), 60% in year 10 (in equal annual increments of 5% over years 6—10), and 80% in year 20 (in equal annual increments of 2% over years 11-20). The Pl,if,t variable then remains constant over years 20-50. This section opened with the supposition that the SCS instant installation assumption would be dropped for expository purposes. Figure 7.4b shows the effect of restoring this assumption, when compared to Figure 7.4a. Only the EB growth function for the Sturgeon is unaffected. Referring to the area under the growth function from time zero to the end of the evaluation period (to year T = 50 or 100, de— pending on the project) as the growthppolygon, the effect of restoring this instant installation assumption is to truncate the growth polygons in Figure 7.4a to year 5 to obtain the polygons in Figure 7.4b (except for that of the Sturgeon). This truncation to year 5 does not remove 250 .sowuoesmmm .mom an msoHum3sm>o uoonoua sasswsuo on» as now: one: mq.n ousMsm as mspouumo nuzonm mm och ecsumssmumos assumes mom was moonsosw nq.n ousmwm use moonsoxm mq.n muswsm .N essay .xwuswns< «monsom Amummmv mass on o . Law a \ \ \ .\\\ .\.L0m .\ \ ‘\ \\\\\\ \ l. \.‘\\ \ 1d -l! \ \ .1 llllll llll.l|l.ll|.ll|lll|lvllllllll\\\ n.“ Anvnoos ”4 \u a o :ommpzum 'o'olo| HHWZ \ I I I l oan \ 10m umpxwoz \ \\ \ \ \ \ \\\ \ A3163 .e.a assess .wuowmopm cmmsnosz snow pom wepwuuwm nusouu mm 251 5 years from the evaluation period. Rather, the effect is to make Pl,if,t in Figure 7.4b equal to Pl,if,t+5 in Figure 7.4a. Essentially, the whole growth polygon is shifted leftward 5 years, and the growth function is extended 5 years. Therefore, the present discounted value of the EB cash flow for any year is increased. The areas under the growth functions in Figure 7.4b are larger than the areas under those in Figure 7.4a (except for the Sturgeon). Sensitivity Analysis Results The EB cash flow achievement rates assumed by SCS will be changed. The achievement rates refer to the vertical distance from the horizontal axis to the EB growth functions, as shown in Figures 7.3 and 7.4, that is, the values of the variable Pl,if,t for various years. Recall that the Pl,if,t data are used only in computing EB, and not in computing FWDRB, nor in computing OTHERB (non-FWDRB and non-EB benefits, excluding secondary benefits). Also the P1,if,t data do not affect the structural costs (SCK and SCOM). Alternatives A and B in Table 7.5 reduce the achievement rates, P1,if,t for all years to 75% and 50% of those assumed in the base estimates. Coincidentally, the NPV sums are reduced to 75% and 50% of the base level, respectively. Alternatives C and D in Table 7.5 discard the SCS instant installa- tion assumption so far as enhancement benefits are concerned. SCK are not affected, although they are in similar assumption changes considered later in this chapter. Alternative C is based on the assumption that the SCS maximum EB rate (maximum P1,if,t) will occur in all years rather than after the completion of the EB development period, usually 15—20 252 Table 7.5. Changed EB Achievement Rates, 12 Michigan PL 566 Projects. NPV sum at 5%, .1 ‘Number of'projecte with Rate change Sludllions "‘IRR's<5%; ""S S B/C's;1.00 Base estimate 13.5 100 1 None (A) SCS rates x .75 10.1 75 3 1 (B) SCS rates x .50 6.7 50 6 4 (C) SCS maximum rate, all years 15.6 116 None None (D) 100% rate, all years 19.9 147 None None Source: computed, program input data, Appendix, Tables 1-3. years. Alternative D is similar except that the SCS partial—achievement rate for BB is discarded in favor of full, 100% achievement in all years (Pl,if,t 8 1.00). As expected, the 12-project NPV sum increased for both alternatives C and D, as compared to the base-estimate NPV. Examination of the EB growth polygons in Figure 7.4b suggests that the impact of alternatives A-D, Table 7.5, will differ by project. Consider the effect of alternatives C and D in Table 7.5 on the EB growth functions for the Sturgeon and Muskrat projects. For assumption alternative C, the Muskrat growth function would be at Pl,if,t = 98% for all years. The Sturgeon function would be at Pl,if,t = 100% for alternative C. The functions for both projects would be at Pl,if,t = 100% for all years in alternative D. The growth polygons become rectangles, whereas with the SCS-assumed P1,if,t achievement rates they were rectangles with the upper lefthand corners cut off to form various , growth patterns. The impact of these alternative EB growth patterns is much.greater for the Sturgeon, because its growth polygon (based on SCS-assumed Pl,if,t data) was less rectangular in shape, compared to the Muskrat's 253 growth polygon. This is revealed in the following NPV data. Alternative C, Alternative D, Base-estimate Table 7.5 Table 7.5 $1,000's ‘ $1,000's ‘ $1,900's Muskrat, NPV at 5% 263 273 281 Sturgeon, NPV at 5% 371 590 590 Muskrat, NPV at 10% 47 53 55 Sturgeon, NPV at 10% -25 83 83 The Sturgeon River project NPV's are considerably increased by the alternative EB farm—income growth patterns, whereas the Muskrat Creek NPV's are affected relatively little. While these two projects were selected because of their clearly divergent base-estimate growth polygons (in Figure 7.4b), other projects are also affected by the postulated growth-pattern changes. This is indicated by the 16% (alternative C) and 47% (alternative D) increases in the 12-project NPV sum (at 5%) in Table 7.5. Incidentally, the effects of alternatives A and B in Table 7.5 can be visualized in Figure 7.4b. Alternative A reduces the Pl,if,t data for all years to 75% of the base-estimate level. Therefore, the heights of all growth polygons in Figure 7.4b are reduced to 75% of the heights shown. The effect of alternative B is similar except that the Pl,if,t data and growth polygon heights for all years are reduced to 50% of the base level. The effect of alternatives ArD in Table 7.6 may be visualized by referring to Figure 7.3. For all alternatives a 20—year, straight—line growth functions is assumed. The SCS-assumed, eventuallysachieved P1,if,t rate occurs in year 20 for all projects. In addition a 20-year 254 associated capital cost (ACK) economic life is assumed. These two changes alone reduced the lZ-project NPV sum (at 5%) to 52%.of the base amount of $13.5 million (alternative A in Table 7.6). For alternatives B-D, Table 7.6, additional changes are introduced. The SCS-assumed, eventually-achieved farm income levels are reduced to 75%, 50% and 25% for alternatives B, C and D, reSpectively. This has the effect of reducing the height of the EB growth polygon in Figure 7.3 (for the straight—line constant rate EB growth pattern); that is, the Pl,if,t data are reduced for all years. As expected, these changes have a sig- nificant impact on the lZ—project NPV sum. Each successive reduction in the achievement rate P1,if,t data of 25% causes the 12-project NPV sum to decline 13%. SCS had been assuming a 50-year economic life for associated capital investments (ACK) in Michigan, but shortened this to 15-30 years, as shown in the Appendix, Table 2 (column 5). Yet, even 15-30 year lives may exceed those apparently recommended by the Agricultural Stabilization and Conservation Service (ASCS, not SCS) in connection Table 7.6. Changed EB Achievement Rates and ACK Economic Lives, 12 Michigan PL 566 Projects Assumption NPV sum at 5%, Number of prpjects with alternatives $ millions IRR's<5% SCS B/C's51.0 Base estimate 13.5 100 1 None (A) No reductiona 7.0 52 5 4 (B) Rates for A_x .75 5.2 39 6 5 (C) Rates for A_x .50 3.5 26 9 8 (D)-Rates for-A x .25. l 7 ~.13w-~ .9 ~ 9 - Source? carpetedfiuagia‘rnput an: 2.53am, Tables 1.33. aAssumes 20—year EB development period and 20-year ACK life only, and no reduction of EB achievement rates Pl if t' 9 9 255 with ACP (Agricultural Conservation Program) payments, for which most, if not all ACK investments would qualify. This information has been available at least since 1957, meaning that it could have been used in all 12 of the studied PL 566 project analyses. Under Michigan conditions an economic life of 15 years may be appropriate for tile drainage systems, which probably account for the bulk of ACK investments.12 Thus, the 20-year economic life assumed for ACK investments in alternatives B-D, Table 6.7, is close to recent SCS practice and to the preceding ASCS recommendations. Cash Flow Timing and Instant Installation As already explained in the methodology section, SCS simplifies the computation of the benefit cost ratios for PL 566 projects by assuming "instant installation." Essentially, this means that investment schedules shown in the SCS project work plans for structural capital cost inflows (SCKt) and land treatment costs (and by implication, associated capital cost inflows, ACKt) are ignored. The SCK capital 12See USDA, SCS, Engineering Division, a letter on the subject: "Life Span of ACP Practices," dated April 25, 1957, from C. J. Francis, Engineering Division director, with recommended life data attached, and with designations from the "1957 National ACP Bulletin." Some of the recommended lives are as follows, although it is not precisely clear from the letter that ASCS recommended them: C—l, permanent sod waterways, 10~15 years; C-4 to C-6, terraces, ditches, dikes and small dams, 10-25 years; C-7, channel linings, chutes, drop Spillways, pipe drOps, etc., 20 years and up; C—8, streambank or shore protection, channel enlargements, floodways, levees, etc., 15-25 years; C-9, open ditches, under 3 feet deep, 5-10 years, or over 3 feet deep, 10920 years; C—lO, underground drains, 15—25 years, with a shorter expected life span applicable if' 'tile is laid in sandy or organic soils, or flatter grades" (underline added); C—12 to C-l6, ‘various irrigation practices, 10—25 years. 256 inflows occurring over the typical 5—year installation period are counted as if they all occurred at time zero. EB achievement rates (Pl,if,t) are advanced in time 5 years. These achievement rates are applied to amortized ACK data in the SCS annual benefit cost ratio, and the year- to-year changes in the rates are assumed to provide an ACK capital inflow schedule for non-SCS investment criteria, as indicated in the methodology section. In this section the SCS simplifying assumption of instant- installation will be discarded. Instead: (1) Structural capital costs (SCK) are allocated over four years using investment schedule data provided in SCS watershed work plans for fiscal and budgeting purposes, but not actually used in project evaluations. (2) Associated capital costs (ACK) are allocated over a period of usually five years, using data provided in SCS watershed work plans for land treatment costs. Land treat- ment costs are similar to ACK, except that they are for the entire watershed area, whereas ACK are for the project- benefited area. Recall that since the SCS evaluations assume only partial completion of the initially-computed EB farm income, this same assumption applies to the ACK amortized inflows, and to the ACK capital inflows for non-SCS investment criteria. Here 100% of the SCS-estimated ACK are assumed to be incurred. (3) The EB farm income levels that SCS assumes will occur at time t = 0 are allocated to occur in equal annual increments over years 1-5. That is, the approach of Figure 7.4a is used rather than that of Figure 7.4b.l3 13An alternative approach to eliminating the instant installation assumption would involve defining the "present" as occurring at year 5 in Figures 7.3 and 7.4a. Net cash.flows would still be counted for years l—T (1—100 or la50, depending on the project), that is, for the full number of years (T) in the project evaluation period. However, net cash flows from time zero to the redefined present (year 5) would be accumulated at interest. Whichever approach is used, the importance of counting all net cash flow components must be kept in mind. Thus, if the redefined "present" (year 5) is used as a point of reference for computing 257 Table 7.7. Effect of Altered Installation Timing, 12 Michigan PL 566 Projects. , _ ..H. .. r Assumption NPV sum at 5%, w A .dNumberlofprojects with alternatives (1 $ millions ' 'IRR'SSS% ' ‘ ' SCS B/C'sil.0 Base estimate 13.5 100 l Noneb (A) SCK onlya 11.8 87 2 —-- (B) SCK and ACKC 8.5 63 5 2 Source: computed, program input data, Appendix, Tables 1—3. aUses main text assumption 1, but not 2; 3 is modified so that the redefined enhancement benefits achievement rates (as in Figure 7.43) apply to all EB components, including ACK. bNot computed. cUses all main text assumptions, 1-3. Introducing each of these assumptions, in the order presented, has the following effects. Allocation of structural capital costs (SCK) to years 1—5 rather than just to year 1 slightly increases the NPV's. As shown in Table 7.8 (Mill Creek project data only), the net cash flows for subsequent years, besides year 1 may become negative, although this is in part due to the delayed EB farm income flows (because of using the patterns of Figure 7.4a, rather than 7.4b). Secondly, use of the full amount of ACK and allocation of ACK over 5 instead of the typical 15 years reduce the NPV's. Thirdly, deferral of SCS—assumed EB achievement rates (in the fashion of Figure 7.4a, instead of 7.4b) reduces NPV's. If one assumes that ACK and land treatment investments are distinct and separate, and that EB growth.patterns properly specify ACK investment rates, then alternative A in Table 7.7 would represent the effect of 13present values, it would be improper to accumulate just SCKt and ACK inflows at interest. That is, by definition net cash.flows for all years also include flows of FWDRB, OTHERB, DNR, FWDC, ACOM AND SCOM. (See discussion in the methodology section.) 258 Table 7.8. Undiscounted Net Cash Flows, Mill Creek Project. fir Yeara Base estimate'. Instant installation-discardedb .$1,000.Ss. ...,... .. ‘$l,000:3i"“""‘ 1 —1,087 -140 2 185 -727 3 201 -11 4 218 -18 5 234 97 6 250 113 7 288 129 8 294 146 9 300 287 10 306 303 11 313 309 12 319 315 13 325 321 14 331 327 15 337 333 16 343 339 17 364 345 18 364 351 19 364 357 20 364 364 21-50 364 364 Source: computed, program input data, Appendix, Tables 1-3. 8Cash flows are discounted as occurring at the beginning of the year; thus, the year 1 flow is discounted as occurring at time zero. bNet cash flows for alternative B, Table 7.7. discarding the SCS instant installation assumption. Alternatively, if one assumes that SCS EB growth patterns refer only to net return achieve- ment, and that they are applied to ACK only as a procedural convenience, then perhaps the land treatment investment patterns shown in SCS work plans represent the agency's better "guesstimate" of investment rates. If this is so, then these land treatment investment patterns should be applied to ACK. In this case, alternative B in Table 7.7 represents the effect of discarding the SCS instant installation assumption. 259 Judging from comments and practices by the SCS Planning Party in Michigan, these land treatment and SCK investment patterns or schedules are shown in work plans only to assist top-level, intra-SCS budget planners to forecast funding requirements to keep various projects on a satisfactory rate of progress. Thus, the schedules have no effect on SCS investment—justifying benefit cost analyses. Concentrating on the explicit land treatment versus implied ACK investment rate patterns (i.e., implied in EB data), the land treatment pattern involves a more complete and rapid investment. This may be compatible with the less complete and slower growth rate for EB net return increases. Recall, SCS may use the EB growth rates on ACK amortized costs (not capital inflows) only as a matter of procedural convenience and simplification. Regardless, how can faster and more complete ACK investment rates be compatible with slower and less complete EB growth rates? The answer would appear to be that SCK and ACK investments elgpe_ are not sufficient to assure full or even partial achievement of with- project farm income levels. These net return benefits credited to a PL 566 project assume: (1) improved farm management, both with respect to successfully operating with newly-drained soils (in Michigan typically) and otherwise; (2) increased usage of fertilizers, pesticides and herbicides, and possibly other inputs in the crop production process; and (3) shifts to creps with.higher net returns per acre, and with (in Mflchigan) lower exceSSemoisture tolerance. Generally, these changes are presumed to represent what is possible with_a given level of farm and crop production technology in the society as a whole, but they do imply a micro or on-farm, indwatershed technological shift. 260 FWDRB Redefined In this section agricultural FWDRB (floodwater damage reduction benefits) will be redefined and counted as an EB (enhancement benefits) component, along with other net farm income. This redefinition is in accord with the concept that the project causes a shift from a lower to a higher net farm income level. For some projects SCS employs this simpler approach, simpler because computation of FWDRB is quite tedious, as explained in chapter 3. SCS personnel have devised procedures that both separate and emphasize FWDRB, at the expense of other forms of farm income. In this author's view, this is as much due to "flood control illusion" as it is to the technical, hydrological rationale for damage estimation. Among proponents of Federal underwriting of flood-control investments, there seems to be the view that FWDRB -- for PL 566, Corps of Engineers, or other projects -- represent loss reductions that are peculiarly in the national interest, whereas drainage and irrigation investments are more private matters which have the disquieting quality of increasing crop production. FWDRB and project investments justified by FWDRB tend to be held less blameful on this account, as discussed in chapter 2. Apart from the separate computation of FWDRB, they are emphasized via EB reductions: EB net income components have associated costs deducted, but FWDRB do not. These reduced EB annual streams do not accrue at their eventuallyeachieved annual rates until a growth.period has elapsed. Even then they do not reach the 100% rate assumed for FWDRB instantaneously (at time zero in the evaluation period). The computation of BB is explained in the methodology section of this chapter for both SCS and non-SCS criteria. The EB component of 261 farm income is DNR (the difference between with.and without project net farm income, DNR = NR2 é NR1, excluding FWDRB). To combine a11.forms of projectecredited net farm income it is necessary to add FWDRB and DNR. By way of explanation, recall that: FWDRB = FWDl - FWD2 (the difference between without and with project floodwater damages) DNR = NR2 - NR1 Redefining DNR as DNR*, we have: DNR* = NR*2 - NR*1 = (NR2 "' FWDZ) '" (NR1 " FWDl) = (NR2 - NR1) -' (FWDZ - FWDl) = (NR2 - NR1) + (FWD1 - FWDZ) = (NR2 - NR1) + FWDRB Using this redefined DRN* in place of DNR significantly affected the NPV's for only 3 of the 12 projects. Each of the remaining 9 projects has unimportant FWDRB. The 12-project NPV sum at 5% was reduced from $13.5 to $11.5 million, or 15%. For projects elsewhere in the country the redefinition could be more important. For the three affected projects: Ratio: [FWDRB / (ave. annual benefits)]14 NPV reduction Cass, S. .26 33% Misteguay .42 21% NMBC (Mill Creek) .29 12% ii a 14Source: Appendix, Table 5. Project average annual benefits exclude secondary benefits. 262 Adverse Farm Income and Capital CosE‘Chapges In this section four kinds of adverse changes are considered and summarized in Table 7.9. Farm income and other benefits (OTHERB) are reduced by application of factors .9, .8, .7, .6 and .5. Recall that project-credited net farm income consists of FWDRB (floodwater damage reduction benefits) and EB farm income. Other benefits (OTHERB) include non—FWDRB and non-EB. However, secondary benefits are excluded in all computations in this chapter. The other three kinds of adverse changes affect capital investment costs. Either associated capital costs (ACK) or structural capital costs (SCK) or both ACK and SCK together are increased by application of factors 1.1, 1.2, 1.3, 1.4 and 1.5. The farm income decreases are more severe in effect than the capital cost increases. A 10% decrease in farm income (and OTHERB) reduced the lZ-project NPV sum to 79% of the base estimate of $13.5 million. This same NPV reduction would require a 50% increase in either ACK or SCK or a 20% increase in ACK and SCK together. As indicated in chapter 6, project-credited farm income estimates are based on a number of crop enterprise assumptions that may be questioned. For example, without government programs, U.S. net farm income would have been 25-50% lower since 1955, according to Tweeten's summary and critique of several economists' studies.15 As shown in Table 7.9, net farm income reductions in the 25-50% range would reduce 15Luther G. Tweeten, "Commodity Programs for Agriculture," in U.S. National Advisory Commission on Food and Fiber, AgriCultural Policy: ’ A Review of'Prggrams and Needs (Washington, D.C.: USGPO, 1967), vol. 5 of the Technical Papers, pp. 107-130. 263 Table 7.9. Comparison of Adverse Net Farm Income and Capital Cost Changes, 12 Michigan PL 566 Projects Change from base NPV sum at 5%, Number ofpprojectvaith estimate data $ millions IRR's<5%‘ SCS BLC51.00 Base estimate $13.5 100% 1 None Farm income and OTHERB are reduced by: (A) 10% $10.6 79% 2 None (B) 20% 7.8 58 5 2 (C) 30% 4.9 36 5 5 (D) 40% 2.0 15 8 8 (E) 50% —0.9 -6 10 10 Associated capital costs (ACK) are increased by: (F) 10% $12.8 95% 2 None (G) 20% 12.0 89 2 None (H) 30% 11.3 84 2 None (I) 40% 10.5 78 2 None (J) 50% 9.8 73 4 2 Structural capital costs (SCK) are increased by: (K) 10% $12.9 96% 2 None (L) 20% 12.4 92 2 None an) 30% 11.8 87 2 1 (N) 40% 11.2 83 4 1 (O) 50% 10.6 79 4 2 Both structural and associated capital costs are increased by: (P) 10% $12.2 90% 2 None (Q) 20% 10.9 80 4 None (R) 30% 9.5 71 5 3 (S) 40% 8.2 61 5 5 (T) 50% 6 9 51 6-w - 5 . SourCef' computed, pngram input data, Eppehaix,‘iah1és 1a3.‘ 264 the lZ-project NPV sum from a base level of $13.5 million to perhaps somewhere in the range of $-0.9 million to $6 million, all Computed.at a 5% discount rate. At higher discount rates the relative decrease in NPV would be greater (see data in Table 7.4 for the effect of a 10% farm income reduction at various discount rates). Regardless of discount rate chosen, it can be appreciated that removing the effect of government farm programs by reducing project—credited net farm income has a signifi— cant impact on project benefits. In chapter 6, use of George Brandow's projected prices, which were intended to remove the effect of government programs, reduced the Mill Creek annual benefit cost ratio from a base of 6.96/l to 2.60/1 (Table 6.8, alternative F). Due to computational differences, the base ratio for this project in chapter 7 is 7.35/1 (Table 7.1). Farm income decreases of 10-50% (Table 7.9, alternatives A-E) reduced the Mill Creek project's SCS B/C from 7.35/1 to the range 6.48/1 (10% decrease in farm income) to 2.97/1 (50% decrease in farm income). The results of the two approaches to removing the effect of government farm programs seem roughly comparable. Given the computational program data inputs used in chapter 7, it is not possible to show the effect of changing crOp prices as in chapter 6. However, the computer program of chapter 7 is far simpler than that of chapter 6.16 16The computer program of chapter 6 is an Operational form of the SCS model described in chapter 3. It uses a host of crop enterprise and watershed data as inputs. The computer program of chapter 7 takes all of this as given and uses DNR, FWDRB, ACK, SCK, and other data as inputs. The DNR and FWDRB data may be viewed as intermediate outputs of the computer program of chapter 6. 265 Besides questioning agency estimates of project.benefits beCause they incorporate the effect of government programs, one may question hydrological, crop yield, crOp cost and other assumptions. All of these assumptions affect FWDRB (floodwater damage reduction benefits) for agricultural areas. Therefore, there may be several reasons for studying the effect of reductions in project-credited net farm income, as in Table 7.9, alternatives A—E. One may also wish to study the effect of capital cost increases (Table 7.9, alternatives F-T) for various reasons. They are used here primarily to show that equivalent percentage decreases in net farm income (on an annual recurring flow basis) and increases in capital costs (on a present value or stock basis) differ in impact. This difference is quite apparent in Table 7.9, regardless of whether one uses the net present value sum, the internal rate of return or the SCS annual benefit cost ratio as a yardstick. Summary This chapter opened with a discussion of various investment criteria. The SCS B/C is the name given to mathematical formulation that provides annual benefit cost ratios approximating those originally computed by SCS. The procedures actually used by SCS are described in chapter 3. Three alternative, non-SCS investment criteria were selected and their mathematical formulations are explained in the methodology section of this chapter. In all three an investment is defined as a negative net cash flow for any year t. In this definition there is no distinction between capital and annual recurring costs. The discounted values of negative net cash flows are summed and counted as the present 266 value of costs. Similarly, positive net cash flows are discounted and summed as the present value Of benefits.‘ These present values are then used to compute present value benefit cost ratios, net present values ' and internal rates of return for the projects. Compared to these other criteria, the SCS annual benefit cost ratio may be conceptually criticized chiefly for the way it incorporates associated capital cost inflows. To be sure, criticism may also be leveled at the agency's discount rates and the way in which they arrange cash flow components in the benefit cost ratio. While there are several problems that may prevent specific conclusions on project rankings, the 12 projects were ranked using selected investment criteria devices. Several of these devices provide similar project rankings: the SCS annual benefit cost ratio (SCS B/C), the internal rate of return (IRR), the present value benefit cost ratio (PV B/C, at 5%) and the single interest rate which, if used in place of the SCS-assumed multiple rates of interest, would provide an SCS B/C = 1. The foregoing project rankings diverge from rankings via net present value data. That is, projects change rank among all criteria, but rank changes are greater in going from the NPV to other investment criteria. There are significant differences in SCS-assumed enhancement benefits (EB) achievement rates for the 12 studied PL 566 projects. Changing the assumed growth patterns affected the computed investment criteria data. Changing both.SCS-assumed EB growth patterns and capital investment rates also affected the investment criteria data. For comparison with chapters 5 and 6, SCS—computed farm income amounts were changed. Also, alternative discount rates were introduced. 267 Given the number of variables, the number of combined effects that could be studied is large. Therefore, few changes were combined. .It may be argued that several of the studied alterations are reasonable alterna- tives to SCS assumptions. Their use in original SCS evaluations would have put some or even all of the projects in a rather bad light. Of course, what is "reasonable" in the way of data and assumptions is Open to discussion and question. Surely, these are matters that public officials may want to consider in determining policy. Keeping in mind the point that what is "reasonable" in the way of assumptions and data is open to question and discussion, perhaps some numerical comparisons may be useful. For the SCS-assumed multiple interest rates, of which there may be as many as four per project, with rate sets differing through time, the lZ-project net present value sum is $19.7 million (Tables 7.1 and 7.4). An NPV sum of $13.5 million is used throughout this chapter as a basis of comparison; it is computed as an across-the—board discount rate of 5%. At a discount rate of 10% the NPV sum is $1.8 million, and at 12% it is $-0.4 million (Table 7.4). Eliminating the effect of government programs could reduce the NPV sum from $13.5 million to somewhere in the range of $—0.9 million to $6 million (Table 7.9), with all NPV's computed at 5%. If the SCS instant installation assumption were discarded and the capital cost investment schedules shown in SCS work plans were applied, the NPV sum would be reduced to $8.5 million from the base level of $13.5 million, with all NPV's computed at 5% (Table 7.7). If in addition SCS—assumed enhancement benefits achievement rates were reduced, 268 further declines in the NPV sum could be.expected, perhaps to somewhere in the $l-5 million range, assuming some combination of resu1ts in Tables 7.5-7.7. Depending on one's assessment of agency assumptions and biases, various combinations of these or other alterations could reflect the effect of removing agency optimism bias in the statement of outcomes. CHAPTER VIII INTEGRATION AND SUMMARY While this chapter is not intended as a summary of individual chapters, some summarization is necessary to integrate the discussion. This will be done in interpretive fashion. The topics include the following: the SCS model, investment criteria, sources of possible error, conclusions, and recommendations. The SCS Model The SCS model described in chapter 3 systematizes SCS procedures for evaluating PL 566 project agricultural benefits. This model is the basis of a computer program used in chapters 5 and 6. Chapter 7 employs a much simpler program and takes agency computed net farm income and other data as given. The mathematical formulations of several investment criteria are shown in chapter 7, including the SCS annual benefit cost ratio. SCS estimates agricultural benefits for PL 566 projects on an annual basis. For policy reasons SCS separates these benefits into FWDRB (flood- water damage reduction benefits) and EB (enhancement benefits). Both relate to the increase in net farm income credited to the project. FWDRB are farm income alone. EB have associated costs deducted. This deduction serves to emphasize FWDRB at the expense of EB. Furthermore, bringing FWDRB into the EB formulation would reduce project benefits, as shown in chapter 7. 269 270 The distinction between FWDRB and EB is computational. Subsequent to their computation, some EB are counted along with.FWDRB to form flood prevention benefits (FPB), as shown in SCS work.plans.‘ Actually, bOth separations relate to policy preferences associated with.what may-be called the "flood control" illusion (discussed below). The following separation of flood prevention and agricultural water management benefits was used for the Mill Creek project: FPB = FWDRB + 1/2 EB = FWDRB + 1/2 MILUB + 1/2 LUCB AWMB - 1/2 EB (AWMB sub-categories are not shown in SCS work plans) The further separation of enhancement benefits is to show the portion for previously cropped land (more intensive land use benefits,‘MILUB) and uncropped land (LUCB, land use change benefits). USDA policy preferences were formulated in 1967: no restrictions were placed on FWDRB, nor on FPB-MILUB; AWMB‘MILUB are not to be based on the increase of output of crops already in surplus. Also, LUCB (both FPB and AWMB-LUCB) are not to be a dominant form Of benefit. All restrictions are discarded for projects planned in Specially designated low-income areas. Benefits and costs are allocated according to dif- ferent rules; therefore, these USDA policy preferences do not have the same meaning as the 1967 House Agricultural Committee policy statement. Flood prevention is to be the unmistakably dominant purpose of projects this committee approves (meaning medium sized projects). The "flood control" illusion mentioned previously relates to the way various project purposes are conceived, regardless of whether the conception is politically or technically motivated. Although all PL 566 project agricultural benefits are based on project—credited increased 271 net farm income, FWDRB are emphasized as being the result of reduced losses, and EB are relegated to the seemingly less desirable role of being increased gains. At the policy level, no question has been raised about FWDRB being in the national interest. Finally the Department of Agriculture did place some restriction on FPB—LUCB, that is, benefits coming from bringing new land into production via flood prevention. Previous to this 1967 restriction, the constraint related only to AWMB-LUCB, that is, benefits coming from bringing new land into crop production via irrigation or drainage. Furthermore, as already indi- cated in the preceding paragraph, FPB-MILUB are accepted, but AWMBAMILUB are disesteemed, unless based on "efficiency" (as contrasted with being based on increasing production of crops already in surplus). The effect of the "flood control" illusion is that farmers whose problems relate to flooding, rather than drainage or irrigation, receive preferential treatment. That is, potential income gains (reduced losses) are allegedly more in the "national interest" for them, while farmers who may be suffering losses in potential income, due to irriga- tion or drainage problems, have potential reduced losses (increased gains) not in the "national interest." Investment Criteria One explanation for possible bias in the estimation of FWDRB may be the policy preference for FWDRB, a preference that relates to the entire federal flood control program, not just to SCS. However, the charge of optimistic bias on the part of agencies concerns the estima— tion of all benefits and costs. One hypothesis is that agencies 272 overstate project net present values (or benefit cost ratios) out of self interest. A counter hypothesis is that economists are overly concerned with some market valued objectives. While the evidence is scanty, several studies suggest that agri— cultural and water resource programs have not been very effective instruments for "improving" income distribution, the most often proposed non-efficiency objective. As a matter of fact, these programs may have "worsened" the income distribution. If this is the case in general, it would appear that agency statements of net present value should be revised downward, rather than upward from their market valued objective base to incorporate the effect of distributional objectives. The three prominent expressions of market valued objectives are equivalent for project set selection with no budget constraint, but not for ranking. Projects may be selected if NPV exceeds 0 or if B/C exceeds 1 at the chosen discount rate or if the IRR exceeds this rate. According to work by Jensen, the net present value is preferable for ranking to either the internal rate of return or the benefit cost ratio because of its responsiveness to changes in investment timing, patterns and scale.1 Such changes are studied in chapter 7, and the NPV is the primary measure of responsiveness. IRR's and benefit cost ratios are used, but only in a qualitative sense. While Jensen employed NPV and 1For the type of changes considered by Jensen, only the NPV is responsive; that is, the computed BIG and IRR data do not change. For the type of changes considered in chapter 7, data computed for all criteria changed, but relative responsiveness may vary according to the criterion. See R. C. Jensen, "Some Characteristics of Investment Criteria," Journal of Agricultural Economics, vol. 20, no. 2, May 1969, pp. 251-268. , I o . ‘ . ' ‘ '_ . 1 a . - ' 7 . .“ 1 1 - 1‘ '_ . o t I t- . "; , u,‘ I ', ' u i 1 . * ‘ r i 7, . I ,- , " " , .1 z . ' ’ . . .“ . ‘, . 1‘ L L .l I I ~ ~ ~ 1 a ' . ~ v 1 1 . - ‘.‘ . a ,1‘ ', ‘o . . II. , '1 , , . - ' . ....n u (. . , u 1 . . . '. ..1. I. J ’,~' .. . ' l. . 1.. ' ‘ I , '. . ' .1'.' . . I ' ) . ‘v .1! . " . , , . , . , 7 . . l ‘ -.1I l. . , '.' J . 7' . 1' l u ‘ ’ - CI ' ; . ’. 5.. . . "I ' , ’ ‘l ' . n .v ' . ‘ . ‘ ‘ ¢.- . I , .I . A 1 1 I. . \ 1 . - . . '. K. I, ( l * .. . I « , i J D '.11 I , , a .,< . . ,.J a 1 . A ‘1 ' 2 ; ’ ‘. u ‘ . " ' t i‘.- " f ' " .‘ut ’ . 5 .. , . ‘ , ' I I 4 ‘ ' . . 1., ', . , . i 1 ' . .. 1 I . I 1 ,1 - .1 _ . , . I ~ - - . a. n . . . ' I I 1' ' ‘l t 1 I} 7': 1 ‘ - 1 ‘ ' , ' i 1 '-‘. I» 7 ‘ ‘. . ‘ ‘. , , . I 1 ( _ ,. ' Q ' . . , .4 ' f I. ' . 1 . ' ‘ . ' "' : . ,1 , , 1 . . , . . . , . o \ . . . . n . . . . A. I . . _. '.::1A .. -1 : r‘ 1 I ’ '. "1 '.1'.' ' . ‘. 1 1'. r 1 . 1 ‘ I ‘ “ I F 4 ‘ . [- .1 ; 1 , '. .- L '._ v I .1 . ' . 1 h _ I ~ , ,‘ .-. , , ' ' . . 1v" ‘. "'- ' ' '1' . , " . 1 - - l , . - . .11 , '..i '1 , l' ‘ . . 15 . '« ' ‘ I . ' \ J I 'I .1” ‘ ’ "A II : ' W1 _ f. f 1. _ .' 1H . II . l . 11' - ' | I. . h ., .' .,, . .. ' ‘ .‘ x .H' .r 1 ;.I ‘ L-‘l ; I W -'"' 1 '.‘ "‘ ‘ ‘ ‘ , V ' . . . .1 u- . . . . . ‘ I 1 . I J. ' ' -f , .~" .11: , _ {I . ‘- '-' ’ ‘5)»: ' 1 3' : 5 ' I "‘H ' I . I:..;( ~ ».- v ‘ ' L . 11’,‘ I" I ' :‘I. . 1,1" 7 I . . F .‘ rm-—: ___- 273 benefit—cost curves to demonstrate the inconsistency of rankings at various discount rates, the author uses NPV curves only, and they suf- fice to explain the inconsistency between the IRR and NPV rankings. In chapter 6 it may have been desirable to use NPV's. However, since the primary concern is with benefit responsiveness, rather than with changes in investment timing, scale and patterns, as in chapter 7, the SCS annual benefit cost ratio is probably an adequate scaling device. With respect to incorporating budget constraints into investment criteria there are several approaches. Eckstein had proposed the use of a benefit cost ratio with Federal cost in the denominator, as discussed in chapter 4. His approach is closest in nature to the agencies' criteria, simply because the agencies use benefit cost ratios. However, the application of his budget constraint device is not as straightforward as may appear at first glance. The agencies' B/C ratios probably are in accord with Senate Document 97, which the agencies authored in 1962. This means that associated costs are counted in the numerator and structural costs in the denominator, as in the SCS benefit cost ratio. Given the data for the 12 Michigan PL 566 projects studied by the author, a B/C ratio with Federal costs only in the denom— inator could be Obtained by shifting SCK to the numerator along with local SCOM, while leaving SCKFederal in the denominator. Besides this, the Federal portion of ACK (paid via ACP payments) should be shifted to the denominator, leaving ACKlocal and ACOM in the numerator. For other agencies this pattern of cost shifts may be inappropriate. The foregoing division of costs ignores the Federal budget impact of loans. If the Federal Government loans to local sponsors and to 274 individual farmer funds to.cover all or a part of their portion of project capital costs (both SCK and ACK), then these loans as well as the outright Federally-paid capital costs constitute demands on the Federal budget. Besides the problem of rearranging cash flows in the benefit cost ratio to accomodate the concept of a Federal budget cost constraint in the denominator, there is the problem of assigning discount rates and timing patterns to various capital inflows. Because of these several difficulties, some computed benefit cost ratios based only on the rearrangement of cash flows are not reported in Chapter 7. The question of discount rates is relevant, because of the possi— bility of Federal loans as well as grants, as indicated previously. If one is going to develop an investment criterion to reflect the difference in risk between publicly borne and privately borne benefits and costs, as suggested by Arrow and Lind (cited in chapter 7), repayment policies and loan terms are of interest. The rationale for this concern is not to establish private borrowing costs, which SCS uses in setting amorti- zation rates for associated capital costs (ACK). As a matter of fact, Eckstein rejects borrowing costs as a basis for multiple rates (as indicated in his discussion of the Greenbook, cited in chapter 7). Rather, the reason for concern with.Federal loan term and repayment policies, as well as with the direct Federal grants (for both SCK and for ACK, via ACP), is that favorable loan policies constitute another element of publicly borne.costs for PL 566 projects. All of this suggests that development of a Federal budget constraint investment criterion is not as straightforward as may seem at first glance, as previously stated. 275 SCS benefit cost ratios represent some sort of mixed budget con- straint. Incidentally, it would seem to the author that there may be some ambiguityin the classification of capital costs between what is counted in the numerator (associated capital costs, ACK) and what is counted in the denominator (structural capital costs, SCK). SCK are presumably for major, mainstream works of improvement, and ACK are for on—farm and inter-farm investments. As suggested in chapter 2, there may be some question about classifying costs for critical-area land treatment practices which are intended to control erosion and runoff above mainstream flood—detention structures. More generally, there may be some question of classification between inter~farm (ACK) and main- stream (SCK) investments. Sources of Possible Error The sources of possible error in the statement of PL 566 project benefit and cost data depend on the assumptions one makes. Certainly consideration should be given to underlying data inputs, procedures and assumptions, as in chapters 5-7. Sensitivity analysis can suggest those items requiring further study. As further studies become available, it will be possible to make some assessment of the degree of error in agency benefit and cost data. FWDRB and Hydrological Assumpiions Compared to the base estimate of $75,715, it was found that certain assumption changes could reduce FWDRB (floodwater damage reduction benefits) for the example project. In the author's judgment this degree of senSitivity is probably characteristic of the storm—to—loss synthe- sizing model, if the assumptions are the same as for the example 276 project. The flood—to—loss synthesizing model was not studied, but it may also have some problematic assumptions. Hydrologists have adapted assumptions used in developing engia neering design criteria for project structural improvements (dams, reservoirs, spillways, channel enlargements and so on) to the estima- tion of FWDRB. FWDRB are computed on the basis of mathematicaly expected hydrological events, and relatively frequent events contribute most to annual expected FWDRB. On the other hand, engineering design criteria are based on less frequent events. Relatively frequent storm rainfall events or flood discharge events are much closer to the critical minimum or threshold level necessary to cause damage; there— fore, they are more sensitive to the assumptions of the storm-to-loss or flood-to—loss translation process. Much of the conservatism applicable in the deve10pment of engineering design criteria is probably incorporated into the estimation of FWDRB, regardless of whether storm or flood event data is used. This conservatism is geared to building structures with a safety factor; that is, capacities are biased on the large side so as to cope with the more unusual combinations of hydro- logical events that lead to severe flooding. Hydrologists indicate that the marginal cost of these safety factors is modest when compared to the initial cost of installing a structure of any size. In estimating agricultural FWDRB a problem arises that is not common to assessing other kinds of flood damages and their aversion. This has to do with the timing of floods during the growing season, for values subject to loss are extremely variable among months and have a range Of 70:1. The SCS model assumes that storms, floods and loss events have the same monthly probabilities of occurrence. The 277 modal month for storms is August, for values subject to loss it is July, but for floods in Michigan it is Mbrchl Even the use Of flood in place of storm monthly probabilities of occurrence could reduce FWDRB perhaps 50-90%, with the reduction being proportional to waterehed size. This does not take account of possible loss avoidance reactions on the part of farmers when floods are most likely (in the months of March, April and May in descending probability order, see Table 5.3). Furthermore, this reduction does not take account of possible upward bias in the estimation of expected acreage flooded, as discussed in the preceding paragraph. Sources of possible error in the estimation of FWDRB for PL 566 projects, as well as those planned by other agencies, are important because of the policy emphasis on FWDRB. As explained in chapter 3 and earlier in this chapter, all agricultural benefits result basically from the project—credited increase in net farm income. Their separation is based on an apparent effort to emphasize FWDRB to conform with this policy preference (see chapter 2). Crop Enterprise Assumptions SCS assumptions for crop prices, costs, yields and planting patterns can be changed by applying adjustment factors to the variable or variables in question. For example, SCS-assumed prices may be increased by 10% for all crops. Another approach is to introduce an alternative set of crOp prices. Both approaches were used in chapter 6. For the first approach, the percentage change in benefits ranged from one to eight times the percentage change in the independent variable. Enhancement benefits (EB) are more sensitive to crop data 278 changes than floodwater damage reduction benefits (FWDRB), because EB include the associated cost deduction, whereas FWDRB are farm income alone. In the author's judgment, SCS use of mid-evaluation period yields overstates project benefits, because a high proportion of the eventually achieved level of farm income occurs at time zero in the evaluation period. FWDRB are based on without-project yields and accrue at their full 100% rate at time zero. EB are based on both with and without project yields, and the annual farm income flows increase during the EB development period, although SCS typically assumes that a high propor- tion of the eventually-achieved annual flow occurs at time zero. In contrast to the author's assertion, it may be argued that use of mid-evaluation period yields to represent the whole period overstates annual farm income flows for the early part of the period and under- states flows for the latter part of the period, assuming an increasing trend of farm income. This is granted. However, the present values of the flows do not balance out. For example, if a $1,000 overstatement occurs for year 1, it is worth $952 at time zero for a discount rate of 5%. A $1,000 understatement for year 50 is worth about $87. The differences in symmetrical pairs of annual flows decrease as one moves toward the middle of the evaluation period, but this does not distract from the fact that using yield assumptions for the middle to represent the entire evaluation period results in an overstatement of the present value of the annual income flows. Government programs have a significant impact on net farm income. Estimates used in chapters 6 and 7 suggest that net farm income would be 25—50% lower without government price and income support programs. 279 The AN (adjusted normalized) crop prices, which have been used by SCS in FL 566 project evaluations since about 1966, are intended to par- tially remove the effect of government programs. For field crops the AN prices are about halfway between the older PLT (projected long term) crop prices, which were used in the period 1957-66, and George Brandow's projected prices. However, for the Mill Creek project, PLT, AN and state average 1959-63 prices and cost adjustment factors all produced about the same level of benefits. For a project with less of the bene- fited area in the watershed planted to vegetable crops, results might differ. A simpler approach to studying the effect of removing government farm programs on project benefits is used in chapter 7. By directly reducing project-credited farm income for 12 Michigan PL 566 projects by decrements of 10 percentage points over the range 10-50%, it was found that the net present value sum (as computed at a discount rate of 5%) could be reduced from a base level of $13.5 million to $10.6 million (for a 10% farm income reduction) or to $-0.9 million (for a 50% farm income reduction). At this lowest extreme only 2 of 12 projects had internal rates of return exceeding 5% (meaning also that their NPV's were positive). Such changes could represent the effect of removing government programs. What about the effect of reducing yield assumptions, reducing or removing FWDRB, or changing crOp planting pattern assumptions? In the author‘s judgment it would not be difficult to produce NPV sums for these 12 projects in the range of say $—10 million to $+30 million, all at a discount rate of 5%. Conceivably this could be done by selecting crop 280 enterprise assumptions and hydrological assumptions in the range over which reasonable men might disagree. It should be added that NPV reductions for farm income decreases are relatively greater at higher rates of discount. For example, as shown in Table 7.4, a 10% farm income reduction decreased the 12 project NPV sum 19% at a discount rate of 5% and decreased the 12 project NPV sum 87% at a discount rate of 10%. For comparison, the 12 project NPV sum is $13.5 million at a discount rate of 5%, and increasing the discount rate to 12% would reduce the sum to $-0.4 million, while decreasing farm income by 50% would reduce the sum to $-0.9 million. In other words, a 50% decrease in farm income has about the same effect as a 140% increase in discount rates. Benefit and Cost Timing, Levels and Patterns, and Discount Rates In terms of present values, changes in the level, timing and pattern of benefit and cost flows, and changes in discount rate may all be significant. Given a set of net cash flow data, with one annual flow for each year in the project evaluation period (of 50 or 100 years), altering the discount rate has the effect of changing the present value of the cash flows. For example, consider flows for the following discount rates, assuming in each.case that the value of the undiscounted annual flow is $1,000, and assuming flows for years 10 and 30: Present value of $1,000~ ~ Relative present value year 10 " year 30 of annual flows r a (0% $1,000? " 77$1,QOQ ' 7 77 7—1/1 ‘ ‘ r = 5% 614 231 3/1 I = 10% 386 57 7/1 r = 15% 247 15 16/1 r = 20% 162 5 32/1 r = 25% 107 l 107/l 281 Thus, not only does increasing the discount rate decrease the present value of future annual flows, but it also changes the relative value of the discounted flows. It is for these reasons that low discount rates promote long-lived projects; that is, at high discount rates, distant—year benefits would have very low present values. At low discount rates distant—year benefits have higher present values in terms of dollar amounts, and they are relatively more important (than they are at high discount rates) compared to nearvyear benefits. The choice of discount rates is, therefore, an important policy decision, not only for project evaluation computations, but for project planning. In other words, if projects are evaluated using relatively high discount rates, the agency will be encouraged to plan shorter-lived projects, simply because distant-year benefits are worth very little in helping to justify projects (i.e., in helping to raise the benefit cost ratio to unity or above). Project economic life and capital intensity are often discussed as being dependent on discount rate and as if they referred to the same thing. This equating may be misleading. It may be related to the simplified (second) presentation of B/Cl’ following. Here, B/C1 is an annual benefit cost ratio and B/Cz is an equivalent present value benefit cost ratio, neither of which is intended to represent the SCS ratio. B/C1 IA.x g (benefits)t.x (1 + r)-tJ / (K.x A‘+ 0M) 1 11 T B/C2 1% (benefits)t;x (l + r)75] / I K.+ ZOMt.x (l + r)7t] B/c1 a B/c2 282 where: A is an amortization factor for economic life T and discount rate r. K represents capital cost. OM represents annual recurring operation-maintenance costs. r is the discount rate. (1 + r)"t is the discount factor, equal to l/(l + r)t T is the project's economic life. Usually B/C1 is stated as follows: B/Cl = (average annual benefits) / (K x A.+ OM) In the second presentation of BIG the benefit summation process is 1 not shown, and it is possible to lose sight of this process and to concentrate on the capital intensity ratio (K/OM ratio) in the denomina- tor. By concentrating on the capital intensity ratio (K/OM ratio), it would appear as if the only effect of changing discount rates is to change the relative importance of amortized capital costs (K x A) as compared to annual recurring operation-maintenance costs (OM). That is, in B/C1 attention is drawn to the ratio (K x A/OM), if the numerator of the B/Cl formulation is taken simply as "average annual benefits," or as "B" in B/Cl = B / (K x A + OM). However, the first formulation of B/Cl, preceding, shows that discount rate does have an effect on the numerator of the B/C ratio. The equivalent present value benefit cost ratio B/CZ’ preceding, focuses even more attention on the variables being considered. That is, there are annual benefit flows, (benefits)t, annual recurring costs, QMt (recurring in the sense that they are usually assumed to be constant and in the sense that theyoccur in all years), one—time capital costs, K, a discount rate, r, and an assumed economic life, T. Changing the 283 discount rate affects the impact of theSe various cash flows in deter— mining the B/C ratio. Cash flows are summed over the range t = l, ..., T, in which T is usually intended to mean project economic life in years. For most PL 566 projects T is 50 years, and beyond this cutoff cash flows are ignored in the summation process. However, the preceding comparison of discounted cash flows of $1,000 for years 10 and 30 shows that at relatively high interest rates distant-year flows are worth very little in helping to justify a project (i.e., in helping to raise its benefit cost ratio), even though cash flows are counted to year T. For example, at 5%, flows beyond year 30 are worth less than $231, decreasing in value with time; at 25%, flows beyond year 30 are worth less than $1, adding virtually nothing to the present value sum, even though their undiscounted value is $1,000. Mathematical formulations for the SCS annual benefit cost ratio and non-SCS investment criteria were presented in chapter 7. Simply stated, the SCS annual benefit cost ratio, SCS B/C, may be presented as follows: SCS B/C (average annual benefits) / (SCK x A + SCOM) 4 (FWDRB + EB + OTHERB) / (SCK x A4 + SCOM) Since FWDRB, OTHERB and SCOM are constants for all years, t = 1, ..., T, this simplified formulation is acceptable for them, However, EB are derived from an amortized present value sum of annual flows and this simplified formulation is misleading. EB = A3zx I? g (DNR - ACK .x A — ACK. x A — ACOM t if if 2 _ 1£,1 1 if,2 _ c—1 , FWDCif) x Pl,if,t / (l‘+ r3) ] if 284 The principal weakness of this definition of EB is that significant capital costs are treated as if they were ordinary annual recurring flows. The noneSCS investment criteria used in chapter 7 compute ACKt as capital inflows rather than as amortized inflows, using the mathemat— ical formulation shown in the methodology section of chapter 7. In these non—SCS investment criteria the capital inflow pattern is based on the EB achievement rate, P1,if,t’ which increases during the EB development period, reaching the SCS-assumed maximum typically by year 20, after which Pl,if,t remains constant. The author assumed that these achievement rates actually refer to EB farm income (DNR), but that applying them to ACK data to Obtain ACKt flows probably represents what SCS might do. Even so, capital investment rates are not actually Specified in the SCS project evaluation procedures, except by implication. However, agency personnel do allocate structural capital costs (SCK) and land treatment costs over a period of a few years for budget planning. The capital investment patterns are shown in SCS watershed work plans, and those for SCK may be adopted directly for use in non—SCS investment criteria. The author assumed that the land treatment investment patterns could be applied to associated capital costs (ACK), which are similar in nature, except that they are for the smaller, projects benefited area, whereas land treatment costs are for the entire waters shed area. In addition, EB cash flow achievement rates were A changed so as to discard the instant installation assumption in their domain. Thus, for one alteration from the SCS assumptions, involving actual specification of both ACK and SCK investment schedules and a delayed pattern of EB farm income achievement, the net present value 285 sum for the 12 studied PL 566 projects fell to $8.5 million, compared to a base estimate sum of $13.5 million, both computed at a discount rate of 5% (as shown in Table 7.7). Without altering the capital investment rates implicitly assumed by SCS, it was found that changing the EB cash flow achievement rates for EB farm income only (DNR) could significantly affect the lZ—project NPV sum. Because these achievement rates differ among projects, specified changes from those assumed in SCS evaluations will vary in impact. As previously indicated, the EB annual flow rates increase during the EB development period, and reach the SCS—assumed maximum typically at about year 20 in the evaluation period (of 50 or 100 years, depending upon the project), after which they remain constant (see Figure 7.3). In one alteration of the SCS-assumed EB cash flow achievement rates, the rates for all years, (t = l, ..., T) were simply increased or decreased. In another alteration a 20—year straight-line development pattern was assumed (see Figure 7.3), with and without individual annual achievement rate changes. The results are presented in Tables 7.5 and 7.6 and will not be repeated here. Suffice it to say that NPV sums ranging from $1.7 to $19.9 million resulted for the 12 studied PL 566 projects, compared to the base estimate NPV of $13.5 million, with all NPV's computed at a discount rate of 5%. Increasing the capital investments by as much as 50% did not pro— duce such a significant change in NPV's. With only ACK increased 50% above the SCS estimated amount the 12vproject NPV sum fell to $9.8 million. With only SCK increased 50% above the SCS estimated amounts, the NPV sum fell to $10.6 million. Even with both ACK and SCK 286 investments for all 12 projects increased to 50% above the SCS estimated amounts, NPV fell to only $6.9 million. 'Again, all NPV's are computed at 5%, and the base estimate NPV sum is $13.5 million. These capital investment changes in effect change the capital intensity ratios for all projects Inot just the simplified ratios, K/OM, suggesting concern only with the denominator of the benefit cost criterion, but the (SCK + ACK) / (SCOM + ACOM) ratios]. Conclusions With reference to hypothesis 1 posed in chapter 1, the author's results show that data computed for various investment criteria are sensitive to underlying assumptions. With respect to hypothesis 2 of chapter 1, if one is willing to specify values for these underlying variables that differ from those assumed by SCS, it would be possible to considerably improve or worsen the apparent worth of the 12 studied PL 566 projects. If the SCS-assumed interest rates (which number as many as four per project and which differ among projects) are applied so far as possible to the various cash flows in a manner that may be assumed to represent what SCS might do, the 12 project NPV sum is about $20 million (i.e., this is the sum of the SCS NPV data for the 12 projects, as shown in Table 7.1). However, in the author's judgment it would be possible to produce a 12 project NPV sum anywhere in the range of say $—20 million to $+40 million, simply by selecting an appropriate combination of alternate assumptions. The reader may ask: After studying SCS procedures and assumptions, could you not specify some alternate assumptions that would give me some idea of what these projects are worth? This is a difficult question to 287 answer with present knowledge. Recall that the author has not actually examined the effects of these projects, even though many of them are now completed and operational. If this had been done, it would provide some appreciation of fairly immediate impacts only, whereas these projects are assumed to have an economic life of 50 or 100 years. Because of the higher degree of sensitivity of investment data to changes in project- credited farm income, what happens in the uncertain future is quite important to any assessment of agency evaluations. In attempting to assess agency evaluations, the problem of assessments being interpreted as critical attacks arises. That is, although the author has assumed that variables and procedures used by SCS in project evaluations represent in essence an expression of agency policy, by virtue of the fact that project evaluations are submitted to an internal agency review process and thereby have the tacit approval of reviewers at the regional level (or even the national level in some cases), there is often some element of innovation in many projects. This type of innovation may come forth at the basic planning level, that is, for example, within the SCS Planning Party in Michigan. However, the following assessments are not intended as critical attacks, but only to indicate the variables where data is important for project selection. To reiterate, it is assumed that variables and assumptions used in project evaluations represent agency policy. And, incidental to this point, it is also assumed that any errors of compu- tation in agency evaluations are of minor importance, as discussed in chapter 1. Disregarding any-other problems, Table 7.1 shows that only 3 of the 12 projects have an internal rate of return exceeding 10%, at which 288 the 12—project net present value sum is $1.8 million. Altering assumptions other than discount rate could reduce the NPV to this level, and if combined with higher discount rates than were Used by SCS, the NPV would be negative. It should be remembered that the problem addressed in this study is the effect of changes in physical and economic variables related to market valued objectives. Even if further study should indicate that changes in data and procedures are warranted and that these changes would reduce the NPV of the 12 projects closer to zero, it still does not answer the broader question of whether the projects are justified on other grounds, such as income distribution or non-market conservation values. Recommendations Given the scope of this study, some recommendations can be made. However, review of the sensitivity analysis results is necessary, as is some sense of perspective for the program as a whole. Program Objectives PL 566 projects relate to market oriented, conservation, develop- ment, flood control (flood prevention), irrigation, drainage and other purposes, objectives and goals. However, improved clarity and under- standing in this matter would be.he1pful. ‘Agricultural FWDRB (floodwater damage reduction benefits) are perceived as being peculiarly in the national interest, but they are in effect just one kind of farm income credited to a project in an agricultural area. Congressional and agency policy statements do not recognize that these effects on income come largely by way of increased output. Yet, there is some constraint 289 on planning and building projects, if the increased income is associated with drainage, irrigation, increased output of surplus crops or increased acreage in agricultural usage. If increased farm income is the objective of this program, then it would be useful to know the costs and benefits of alternative means of accomplishing this objective. If flood loss reduction is the objective of the PL 566 program, then information on alternative means of achieving this objective would be helpful, hopefully including means that would not at the same time increase agricultural output. For example, shifts from crop to pasture usage of farm land in frequently flooded areas would reduce flood damages, but farm income would probably also be reduced. By contrast, projects reduce damages by decreasing flood hazards rather than by decreasing the values subject to loss. Sources of Possible Error in FWDRB Estimates FWDRB (floodwater damage reduction benefits) are the most important category of benefits for the PL 566 program nationally (Appendix, Table 5). They are obtained as the difference between without and with project damages which are computed in accordance with the concept of mathematical expectation (see chapters 3 and 5). Consequently, relatively frequent events account for the bulk of expected damages. Because of this, typical acre loss values (composite acre values, CAV) for say the lOalS year flood zone Could be used to estimate FWDRB, rather than loss values based on the assumption that cr0pping patterns and land use are homogeneous throughout an entire ecOnomic reach (including area beyond even the flood zone for the largest flood used to estimate FWDRB, usually a 50-year or lOO-year flood). The use of loss 290 values for the smaller area would be.hased on the assumption that farm managers change land use in frequently flooded areas as a loss-avoidance reaction. I Another source of possible error in the estimation of FWDRB has to do with the use of storm instead of flood data in forming sets of monthly probabilities of loss. The modal month for storm probabilities is August, and that for floods is March.in Michigan; use of flood in place of storm data would considerably reduce FWDRB (Table 5.3). Furthermore, there are several technical hydrological assumptions that are part of the process of translating storm into flood magnitudes, and changes in any one of these could significantly affect FWDRB (chapter 5). Even a seemingly minor error in estimating the runoff curve number could change FWDRB by 150% (PP. 146-147). Changes in crop enterprise assumptions could significantly affect the relative importance of FWDRB (see chapter 6, especially the relative benefit data in Table 6.1, p. 186 and that in the informal table, p. 207); so would separation of associated costs from enhancement benefits (see Tables 3.7 and 3.9 and chapter 7). Sensitivity of Benefits to Farm Income Assumptions The several possibilities of error.mdght lead one to discard Iagricultural FWDRB estimates (floodwater damage reduction benefits estimates). Thus, one could simply compute project benefits on the basis of the difference between with and without project net farm income (an alternative sometimes used by SCS, see Table 3.1 and related discussion in chapter 3). This was done for the Mill Creek project. Reducing 291 without—project (flood-free) yield levels to a with—flooding level gave the same benefit cost ratio, as explained in chapter 6 (see pp. 193el94). That is, the without-project farm.income level was reduced so that the project-credited amount of farm income increased. Essentially, project benefits are computed in the enhancement benefits (EB) sub-routines when FWDRB are not estimated. While FWDRB deserve Special attention because they constitute over half of the benefits used to justify PL 566 projects for the nation as a whole (see Appendix, Table 5 for data), their elimination would not entirely remove agency estimates from the realm of doubt as to the possibilities of error. As explained in chapter 6, all agricultural project benefits depend upon certain crop enterprise assumptions, namely crop prices, production costs, yields and planting patterns. EB (enhancement benefits) are generally more reaponsive to changes in these variables than are FWDRB, because EB are partially net benefits (project-credited farm income with associated capital costs deducted), while FWDRB are project-credited farm income alone. Benefit response coefficients for these variables are shown in Table 6.2. It would be useful to compare the sensitivity of these crop enterprise variables with the sensitivity of hydrological variables. Unfortunately, changes in hydrological variables, as studied in chapter 5, can not be readily expressed in percentage terms; therefore, benefit response coefficients can not be used to express the results of chapter 5. The picture of relative importance of different variables is further clouded by the fact that not even all of the Crop enterprise variable changes in chapter 6 lend themselves to expression in terms 292 of benefit response coefficients." In addition to crop enterprise and hydrological variables, it would be useful to compare the effects of changes in things like investment criteria, discount rates, and benefit and cost flow rates. This is difficult because the benefit response information in chapter 6 is not readily comparable with the effects of alternative investment criteria variables in chapter 7; that is, benefits, the numerator of the benefit cost ratio as defined by SCS, are used as the primary measure of reSponse in chapter 6, whereas net present value is used in chapter 7. It is inherently difficult to speak of a given unit of change in these various kinds of estimation components. Still, some comparisons are possible, but they require an appreciation of the relationship between the SCS annual benefit cost ratio and the net present value (see the detailed presentation in the methodology section of chapter 7). For purposes of rough comparison, if the SCS annual benefit cost ratio is unity (benefits / costs = l), the net present value is zero. It should also be kept in mind that the benefit cost ratio data in chapters 5 and 6 are for one project, while the net present value sums in chapter 7 are for 12 projects. To reduce the lZ-project NPV sum to zero would require an interest rate of about 11% [120% above or 2.2 times the 5% rate used to obtain the base estimate NPV sum of $13.5.million (Table 7.4)]. Given the assumptions on cash flows used by SCS (chapter 7), perhaps a 100% increase in capital costs would also reduce the lanroject NPV sum to zero (at 5% interest, Table 7.9). At 5% interest, a 50% decrease in project-credited farm income would reduce the NPV sum to zero (Table 7.9), although at higher interest rates NPV-sum responsiveness is 293 greater (Table 7.4). Assuming that a lZ—project NPV sum of zero is roughly comparable to an SCS annual benefit cost ratio of unity, any of the following adverse changes in crop enterprise variables could reduce the NPV sum to zero: (1) a 5-25% reduction in withqproject yields (yield set 2), (2) a 12—35% increase in without-project yields (yield set 1), (3) a 25% decrease in crop prices, or (4) a 35-45% increase in crop production costs (p. 192). Simultaneous reduction of about 40% in both yield sets might reduce the ratio to SCS B/C = l, or make NPV = 0. Extension of conclusions would require consideration of the importance of benefit categories (Table 6.1), because the example projects' responsiveness of benefits (Table 6.2) relates to the responsiveness of enhancement benefits which dominate (comparing the sum of MILUB and LUCB to FWDRB in Table 6.1). Specifically, reSponsiveness of FWDRB to changes in crop enterprise variables should be noted, since FWDRB account for a larger portion of total benefits for PL 566 projects nationally than they do in Michigan (Appendix, Table 5). Suppose that a project were justified only by FWDRB, then comparisons between the results in chapters 5-7 may become more meaningful. In chapters 5 and 6, the Mill Creek project would have a benefit cost ratio of about 1.9/1 (FWDRB I annual costs a $75,715 / $40,520 a 1.9, Table 6.1), and reduction in benefits of about 50% would reduce the ratio to SCS B/C = l, or make NPV = 0.. This 50% reduction in FWDRB.would be the result of any of the following changes: a 25% decrease in crop prices, a 40% increase in crop costs, or a 60% decrease in without-project yields (yield set 1, Table 6.3). 294 FWDRB reSponsiveness differs for the Tebo Erickson project, and a 50% decrease in FWDRB would result from the following respective changes in these crop enterprise variables: prices, ~30Z; costs, +70%; and without-project yields, —50% (Table 6.4). Thus, a re-ordering of the impact of crOp enterprise variables is necessary if FWDRB dominate, because of differing benefit reSponsiveness. For overall benefits, for the example projects in chapter 6, the ordering of adverse crop enterprise variable changes to make SCS B/C = l is as follows: with-project yields, without-project yields, crop prices and crop costs or simultaneous yield set changes. For FWDRB alone the ordering is prices, costs and without-project yields. Any one of a number of changes in hydrological assumptions or variables could reduce FWDRB by 50% (see chapter 5, summary). For example, (1) a slight field error in the estimation of the runoff curve number, (2) a shift to by-month runoff curve estimation (Table 5.4), (3) use of the point-area rainfall correction and annual~maximum series rather than point-estimate partial-duration series storm rainfall data, (4) a shift from storm to flood data as the basis for developing monthly probabilities of flood loss (Table 5.3), or (5) some combination of these changes. Therefore, hydrological assumptions are critical in the justification of projects where FWDRB dominate, as they do for the nation as a whole (Appendix, Table 5). However, if these hypothetical changes should later be discovered to be factual by further research, there remain some serious questions about the impact of crop enterprise variables as a source of possible error. 295 Sensitivity of Project Worth to Benefit and Cost Timing, Pattern andBatefAssumptions Given the computed farm income amounts, they must be further specified as cash flow rates for all years in the evaluation period. Also, capital investment rates must be specified, since PL 566 projects involve multi-period rather than single-time (time zero) capital investments. SCS procedures involve some simplification in these matters, but specification of various cash flow rates is preferable. Changes in various cash flow rates are studied in chapter 7. In the author's judgment, cash flow rates should be separately specified for project-credited farm income (which may consist of several components, with each component requiring a specified cash flow rate), associated capital cost investments (ACK) and structural capital cost investments (SCK). Of course, other kinds of benefits and costs may also require‘ the specification of cash flow rates. The results in chapter 7 suggest several possibilities of error in actual project evaluations. Some of these results are based on the use of cash flow rates estimated by SCS, but not used by SCS in the project benefit cost evaluations (Tables 7.5-7.7). The effect of one alteration from SCS assumptions is shown in Table 7.8, which presents a comparison of annual net cash flows. . Recommendationsffor Further Study ..Va‘jwfi Based on the preceding discussion of variable sensitivity, some recommendations for further study can be made. 296 (1) Assuming FWDRB (floodwater damage reduction benefits) are to continue to dominate total annual benefits for PL 566 projects as they have in the past, underlying hydrological variables should be given detailed consideration from the standpoint of FWDRB estimation. These variables have been studied by hydrologists primarily from the stand— point of developing engineering design criteria, but the flood discharge and runoff amounts used to estimate FWDRB are much closer to the critical minimum or theshold level just necessary to cause overbank flow onto the floodplain. Studies in other areas of the country are suggested because hydrological conditions may vary. Economists should not become obsessed with arguments over things like the discount rate and fail to see the critical sensitivity of benefit estimates to small changes in hydrological variables. (2) Still assuming FWDRB dominance, there are a host of non- hydrological variables that deserve study. Even if the FWDRB procedure is discarded because of its high sensitivity to hydrological variables, for which refined estimates may be found to be simply not possible or too costly to obtain, many of these other variables remain to raise questions about the possibilities of error in the statement of project worths. Generally, variables affecting the estimated cash flows of farm income are more important than variables affecting capital cost flows and more important than discount rates. (a) A major policy issue concerns whether projects should continue to be justified on the basis of crop prices and related input cost adjustment factors that are based in part on government farm.price and income support programs. Water Resources Council price data do 297 not remove all of the effects of these programs. This is an important policy choice because of the Sensitivity of benefit estimates to crop prices. (b) Mid-evaluation period yields essentially overstate project worth. The degree of bias depends on the interest rate chosen to discount future income streams. The higher the interest rate, the greater the degree of overstatement, since distant-year income streams, which would be understated (in terms of undiscounted values) by use of mid-evaluation period yields, are worth less and less as the interest rate used for discounting is increased. Therefore, the overstatements for years before the mid year are not balanced out, except at zero discount rate. Yield assumptions should be studied with this in mind. (c) Cropping patterns can affect apparent project worth. It is recommended that study be given to the cropping and land use patterns in the 10-15 year flood zone. If they are different than those in the surrounding area, it may be argued that they should be used to estimate FWDRB. (d) Enhancement benefits achievement rates for farm income should be given separate consideration from capital cost investment rates. All of them can affect project worth. In this connection, the SCS instant-installation assumption and the use of amortized rather than capital cost flow rates.may be viewed as computation simplifying assumptions that distract attention from important variables in the project analyses. BIBLIOGRAPHY BIBLIOGRAPHY Adams, Dale W. The Benefit-Cost Analysis - Its Key Variables - and Public Law 566 in Michigan. Agricultural Economic Department Publication No. 904. East Lansing, Michigan: Department of Agricultural Economics, Michigan State University, March, 1963. . "Benefit-Cost Analysis on Public Law 566 Projects in Michigan." Unpublished M.S. thesis. East Lansing, Michigan: Department of Agricultural Economics, Michigan State University, 1961. Alchian, Armen A. "The Rate of Interest, Fisher's Rate of Return Over Costs and Keynes' Internal Rate of Return," American Economic Review. Vol. 45 (December, 1955), pp. 938-943. Arrow, Kenneth J. "Criteria for Social Investment," Water Resources Research. Vol. 1, No. 1 (First Quarter, 1965), pp. 1-8. Arrow, Kenneth J. and Lind, Robert C. "Uncertainty and the Evalua- tion of Public Investment Decisions," American Economic Review. Vol. 60, No. 3 (June, 1970), pp. 364-378. Baumol, William J. "On the Appropriate Discount Rate for Evaluation of Public Projects," Congressional Record - Senate. (September 26, 1967), pp. 813692-813696. "On the Social Discount Rate,” American Economic Review. Vol. 58 (September, 1968), pp. 788-802. Bell, Frederick C. Estimating Design Floods from Extreme Rainfall. Hydrology Papers, No. 29. Fort Collins, Colorado: Colorado State University, July, 1968. Benson, M. A. "Plotting Positions and Economics of Engineering Planning," Proceedings of the American Society of Civil Engineers, Journal of the Hydraulics Division. Vol. 88, No. HY6, Part 1 (November, 1962), pp. 57-61. Bonnen, James T. "Rural Poverty: Programs and Problems," Journal of Farm Economics. Vol. 48, No. 2 (May, 1966), pp. 452-465. 298 299 Castle, Emery; Kelso, Maurice and Gardner, Delworth. "Water Resource Development: A Review of the New Federal Evalua- tion Procedures," Journal of Farm Economics. Vol. 45, No. 4 (November, 1963), pp. 693-704. Chase, Samuel B., Jr. (editor). Problems in Public Expenditure Analysis. Washington, D. C.: The Brookings Institution, 1968. Chow, Ven Te. Handbook of Hydrology: A Compendium of Water Resource Technology. New York: Mc Graw-Hill Book Co., 1964. Dixon, Wilfred J. and Massey, Frank J., Jr. Introduction to Statis- tical Analysis. 2nd ed. New York: NC Graw-Hill Book Co., Inc., 1957. Eckstein, Otto. Water Resource Development. Cambridge: Harvard University Press, 1958. Feldstein, M. S. "The Social Time Preference Discount Rate in Cost Benefit Analysis," Economic Journal. Vol. 64 (June, 1964), pp. 360-379. Ford, Erwin C., Cowan, Woody L. and Holtan, H. N. "Floods--and a Program to Alleviate Them," Water: The Yearbook of Agri- culture, 1955. U.S. Department of Agriculture. Washington, D. C.: Government Printing Office, 1955. "The 4-1/4 Per Cent Rate Ceiling, Blessing or Curse," Monthly Economic Letter. New York: First National City Bank, May, 1969, pp. 52-54. Freeman, A. Myrick, III. "Adjusted Benefit-Cost Ratios for Six Reclamation Projects," Journal of Farm Economics. Vol. 48, No. 4, Part I (November, 1966), pp. 1002-1012. "Six Federal Reclamation Projects and The Distribution of Income," Water Resource Research. Vol. 3, No. 2 (Second Quarter, 1967), pp. 319-332. Fuller, Vernon. "Political Pressure and Income Distribution in Agriculture," Journal of Farm Economics. Vol. 47, No. 5 (December, 1965), pp. 1245-1251. Hardin, Charles M. Food and Fiber in the Nation's Politics. Vol. III of Technical Papers. National Advisory Commission on Food and Fiber. Washington, D. C.: Government Printing Office, 1967. Haveman, Robert H. "The Opportunity Cost of DiSplaced Private Spending and the Social Discount Rate," Water Resources Research. Vol. 5, No. 5 (October, 1969), pp. 947-957. I -- 44' '~ 1. I 300 Haveman, Robert H. Water Resource Investment and the Public Interest. Nashville: Vanderbilt University Press, 1965. Held, R. Bernell and Clawson, Marion. Soil Conservation In PersPective. Baltimore: Johns Hopkins Press for Resources for the Future, 1965. Jensen, R. C. "Some Characteristics of Investment Criteria," Journal of Agricultural Economics. Vol. 20, No. 2 (May, 1969), pp. 251-268. Kamien, Morton I. "Interest Rate Guidelines for Federal Decision- making," Congressional Record - Senate. (September 25, 1967), pp. 813543—813545. Krutilla, John V. and Eckstein, Otto. Multiple Purpose River Develop- ment. Baltimore: Johns Hopkins Press for Resources for the Future, 1958. Kuiper, Edward. Water Resource Development--Planning, Engineering and Economics. London: Butterworth and Co., Ltd., 1965. Leopold, Luna B. and Maddock, Thomas, Jr. The Flood Control Contro- versy. New York: The Ronald Press, 1954. Maass, Arthur. "Benefit Cost Analysis: Its Relevence to Public Investment Decisions," Quarterly Journal of Economics. Vol. 80, No. 2 (May, 1966), pp. 208-226. Marty, Robert. "The Composite Internal Rate of Return." Unpublished draft copy. East Lansing, Michigan: Department of Forestry, Michigan State University, 1970. McKean, Roland. Efficiency in Government Through Systems Analysis. New York: John Wiley and Sons, Inc., 1958. Mead, Daniel W. Hydrology--The Fundamental Basis of Hydraulic Engineering. 2nd ed. revised. New York: McGraw-Hill Book Co., 1950. Michigan, Water Resources Commission. Hydrological Studies of Small Watersheds in Agricultural Areas in Southern Michigan. Reports 1, 2, and 3. Lansing, Michigan: Water Resources Commission, 1958, 1960, and 1968. Moreell, Ben. Our Nation's Water Resources--Policies and Politics. Chicago, Illinois: The Law School of the University of Chicago, 1956. Morgan, Robert J. Governing Soil Conservation. Baltimore: Johns Hopkins Press for Resources for the Future, 1965. 301 Myers, Earl A. and Kidder, E. H. "Practical Hydrological Concepts Concerning a Small, Michigan, Agricultural Watershed," QuarterlyyBulletin. Vol. 43, No. 4. East Lansing, Michigan: Michigan Agricultural Experiment Station, Michigan State University, May, 1960, pp. 743-750. Ogrosky, Harold O. "Hydrology of Spillway Design: Small Structures-- Limited Data," Journal of the Hydraulics Division, ASCE (American Society of Civil Engineers). Vol. 90, No. HY3, Proceedings Paper 3914 (May, 1964), pp. 295-310. Prest, A. R. and Turvey, R. "Cost-Benefit Analysis: A Survey," The Economic Journal. Vol. 65, No. 300 (December, 1965), pp. 683-7350 Robinson, Kenneth L. "The Impact of Government Price and Income Programs on Income Distribution in Agriculture," Journal of Farm Economics. Vol. 47, No. 5 (December, 1965), pp. 1225- 1234. Ruttan, Vernon W. The Economic Demand for Irrigated Acreage. Baltimore: Johns Hopkins Press for Resources for the Future, 1965. Smith, Stephen C. and Castle, Emery N. (editors). Economics and Public Policy in Water Resources Development. Ames, Iowa: Iowa State University Press, 1964. Spurr, William A., Kellogg, Lester S. and Smith, John H. Business and Economic Statistics. Revised ed. Homewood, Illinois: Richard D. Irwin, Co., 1961. Stockfisch, J. A. "The Interest Rate Applicable to Government Investment Projects," Congressional Record - Senate. (September 22, 1967), pp. 813467-813472. Tinsley, W. A. Rates for Custom Work in Michigan. Extension Bulletin E-458, revised. East Lansing, Michigan: Michigan State University, Cooperative Extension Service, February, 1967. Tolley, George S. and Riggs, Fletcher E. (editors). Economics of Watershed Planning. Ames, Iowa: The Iowa State University Press, 1961. Tweeten, Luther G. "Commodity Programs for Agriculture," Agricultural Policy: A Review of Programs and Need. Vol. 5 of the Technical Papers. National Advisory Commission on Food and Fiber. Washington, D. C.: Govern- ment Printing Office, 1967, pp. 107-130. 302 U.S. Army, Corps of Engineers, District Engineer (of Sacramento, California). Statistical Methods in Hydrology. Revised ed. Sacramento, California: U.S. Army, Corps of Engineers, January, 1962. U.S. Congress. Policies, Standards Evaluation and Review of Plans for Use and Development of Water and Related Land Resources. Senate Document 97. 87th Cong., 2d Sess. Washington, D. C.: Government Printing Office, 1962. . The Soil Conservation Service Act (also called The Soil Erosion Act and The Soil Conservation Service Establishing Act). Public Law 46 (49 Statutes 163, 164; 16 U.S.C. 590a- 590f). 74th Cong., l Sess. 1935. References are made to this Act, as amended through 1947. The agency was established as the Soil Erosion Service in 1933 in the Department of Interior under the National Industrial Recovery Act of 1933. Public Law 46 of 1935 renamed the agency as the Soil Conservation Service and transferred it to the Department of Agriculture. . The Watershed Protection and Flood Prevention Act. Public Law 566 (68 Statutes 666). 83d Cong., 2d Sess. 1954. References are made to this Act, as amended through 1968. U.S. Congress, Joint Economics Committee. Economic Policies for Agriculture in the 1960's. 86th Cong., 2d Sess. Washington, D. C.: Government Printing Office, 1960. U.S. Congress, Joint Economics Committee, Subcommittee on Economy in Government. The Analysis and Evaluation of Public Expenditures: The PPB System. Vol. 1, Vol. 3. Wash- ington, D. C.: Government Printing Office, 1969. This collection of papers is referred to as JEC-PPBS in the text. . Hearings on Economic Analysis of Public Investment Decisions: Interest Rate Policy and Discounting Analysis. 90th Cong., 2d Sess. Washington, D. C.: Government Printing Office, 1968. U.S. Department of Agriculture. Agricultural Statistics. Washington, D. C.: Government Printing Office, Years 1960-1968. U.S. Department of Agriculture, Agricultural Research Service and Agricultural Marketing Service. Agricultural Price and Cost Projections for Use in Making Benefit and Cost Analysis of Land and Water Resource Projects and Analyzing the Repayment Capacityyof Water Users. Washington, D. C.: U.S. Department of Agriculture, September, 1957. 303 U.S. Department of Agriculture, Agricultural Stabilization and Conservation Service, State Office in Michigan. Agricultural Conservation Handbook for 1965, Michigan. East Lansing, Michigan: Agricultural Stabilization and Conservation Service, November, 1964. U.S. Department of Agriculture, Economic Research Service. Demand and Price Situation. DPS-115. Washington, D. C.: United States Department of Agriculture, February, 1968. U.S. Department of Agriculture, Economic Research Service, Natural Resource Economic Division. "Inventory of Basic Data - Public Law 566 Watershed Work Plans." October, 1967. U.S. Department of Agriculture, Soil Conservation Service. Economics Guide for Watershed Protection and Flood Prevention. Wash- ington, D. C.: Soil Conservation Service, 1964. This publication is referred to as Economics Guide in the text. . Hydrology Guide for Use in Watershed Planning. By Victor Mockus. Washington, D. C.: Soil Conservation Service, November, 1954. . National Engineering Handbook, Section 4, Hydrology, Part 1: Watershed Planning. By Victor Mockus. Washington, D. C.: Soil Conservation Service, August, 1964. This publication is referred to as either NEH-4 or The Hydrolpgy Handbook in the text. . SCS National Engineering Handbook, Section 4, Supple- ment A. By Victor Mockus. Washington, D. C.: Soil Conservation Service, 1956. This publication is referred to as NEH-4 (1956 ed.) or The Hydrology Handbook (1956 ed.) in the text. . Watershed Protection Handbook, Part 1: Planning and Operations. Washington, D. C.: Soil Conservation Service, August, 1967. This publication is referred to as WPH in the text. U.S. Department of Agriculture, Soil Conservation Service, Engineer- ing Division. "Life Span of ACP Practices." April 25, 1957. This is a letter on the subject by C. J. Francis, Engineering Division director. U.S. Department of Agriculture, Soil Conservation Service, Engineer- ing and Watershed Planning Unit. "Evaluating Flood Damages to Crops and Pasture," Memorandum No. 3 Revised. October 23, 1958. 304 U.S. Department of Agriculture, Soil Conservation Service, Engineering and Watershed Planning Unit. "Most Damaging Versus Largest Storm Damage," Intra-Agency CorreSpondence. June 1, 1958. This is a letter on the subject from Frank E. Erickson, hydraulic engineer, to Kermit R. Irwin, a member of the SCS Watershed Work Plan Party in Missouri. U.S. Department of Agriculture, Soil Conservation Service, SCS Administrator (D. A. Williams). "Agreement with Corps of Engineers with Respect to Flood Protection by Engineering Works," Watershed Memorandum 75. December 14, 1965. "Federal Assistance in Watershed Projects," SCS Advisory WS-28. November 18, 1965. "Flood Prevention in Watershed Projects," Watershed Memorandum 86. September 28, 1967. Attached is a letter from W. R. Poage, Chairman of the House Agriculture Committee, to Speaker of the House, John W. McCormack, dated July 31, 1967. The author refers to this statement as the 1967 House Agriculture Committee's policy statement on PL 566 project purposes. The state- ment relates to the required dominance of flood prevention in terms of Federal contribution to construction costs for projects approved by this Committee. "New Discount Rates," SCS Advisory WS-l9. August 1, 1968. . "Surplus Crop Production," Watershed Memorandum 84, Supplement 1. August 25, 1967. Attached is a letter from John A. Baker, Assistant Secre- tary of Agriculture for Rural Development and Conservation, to the Secretary of Agriculture (Orville Freeman). The author refers to this statement as the 1967 USDA policy statement on PL 566 project purposes. It is based on benefit data assessments and relates to planning approval. "Watershed Land Treatment," Watershed Memorandum 70. November 5, 1964. "Watershed Protection (PL S66)--Land Treatment Measures in Watershed Work Plans," Advisory Notice W-748. Septem- ber 28, 1962. U.S. Department of Treasury. Treasury Bulletin. Washington, D. C.: Government Printing Office, January, 1970. 305 U.S. Geological Survey. Magnitude and Frequency of Floods in the United States,yPart 4,78t. Lawrence River Basin. Geological Survey Water Supply Paper 1677. By S. W. Wutala. Washington, D. C.: Government Printing Office, 1965. U.S. Office of the President, The Task Force on Federal Flood Control Policy. A Unified National Program for Managing Flood Losses. House Document No. 465, 89th Cong., 2d Sess. Washington, D. C.: Government Printing Office, 1966. U.S. Weather Bureau. Rainfall Frequency Atlas of the United States for Duration from 30 Minutes to 24 Hours and Return Period from 1 to 100 years. Technical Paper No. 40. By David M. Hershfield. Washington, D. C.: Government Printing Office, May, 1961. This publication is referred to as TP-40 in the text. . Rainfall Intensity - Frequency Regime, Part 5 (of 5)-- Great Lakes Region. Technical Paper No. 29. By David M. Hershfield. Washington, D. C.: Government Printing Office, 1960. This publication is referred to as TP-29 in the text. . Two to Ten Day Precipitation for Return Periods of 2 to 100 Years in the Contiguous United States. Technical Paper No. 49. By John F. Miller. Washington, D. C.: Government Printing Office, 1964. This publication is referred to as TP-49 in the text. U.S. Water Resources Council. Interim Price Standards for Planning and Evaluating Water and Land Resources. Washington, D. C.: Water Resources Council, April, 1966. Vondruska, John. Estimating Small Watershed Project Benefits: A Computer Systemization of SCS Procedures. Agricultural Economics Report. Report No. 120. East Lansing, Michigan: Department of Agricultural Economics, Michigan State University, February 1969. Wisler, Chester 0. and Brater, Ernest F. Hydrology. 2d ed. revised. New York: John Wiley and Sons, Inc., 1950. APPENDIX 306 n .Hmpmcomnco: Aaaummmwoo: uo: usn .wccsm mom amino: mcacmwe .mumoo :pwcuo: mo muomoumo map :a cmucsoo was 0mm.mmm wasp .wanmu swan xuoz mum on» :a momm.mmw pom acmpm .monmEEoo mo ucoEuamama .Amm anocoom cm mmusaoCa .Uama maampmcmm .Mom .coowusumn .wquHmm ou oocmumammm AEmawopQ cowum>pwmcoo amHSuasoapwmv mo< ma> pawn haamnmwwa on zme xo< mo Nom mamsumm .muaq mooo3 Enmm pom anqHz pom . HM0< «Eammunouca .N.Maxo< «Eummnco .H.maMo< «Han mmo< “moaa mauaImusummm pom.a.mxo< ”czonm no: mam Mo< mo mcoama>au opem .umm» wcaccmHa can you mmumEaumm wcanwwcawcw co comma mum Awumoo quEumoHu .oanmu on» :a nonmaa mm .muowmopa NH ozu wow .%Hm>wuowamwp mama cam xo< mam Momv mumoo amuammu .oo.a .oo.a .mm. .mm. cam .sm. “mm. .qw. .Nm. mow. .ww. umzoaaom mm mwcmu mm.o ou qw.o wcu ca maam5m: .nouomm xwvcw Acmnaameuoc owumsncmv z< Ho AEpwu mcoa nmuoonopav Ham mnu Hmcuww mo own an wuwo coaumqu55ooo map you mama onu Eoum oumaczoo owumsmcm mum Azoo< cam zoomv mumoo mocmcmuchelcoaumaoao .AHCO COwumquESUOU EOHM wumoo Uwumwoommaa. .GOHumquESUOU USN mCmHQ x903 vwflmhwumas mmom «mam: quHaaom aaq.mmm.a oao.oma.oa omm.oaa.m osm.osa amm osmwoqm omm.m coaummq oqo.mmm omm.mma moma mama ooa.mmm mom.m mmm.ms~ omm.m oms.oom pmam.mma mam.o~a coma commasum mam.omm.a ssm.m mma.mmo.a omm.a Nmm.mom mo~.mom mam.mmm Noma ammao aaaz mom.mm --- sqm.mma oom.a mqm.mm moo.mm mam.om mmma umaxmsz mmm.~ma.m omo.m~ cms.moq.m moa.aa 6mm.omm qqm.amm Nmm.mms mmma mmswmumaz omm.sm mam.~ mao.oaa om~.N own.ama omm.maa oom.mm Nmma maauaa mmm.mm~ mos sma.~ma Nam.a ooh.ma~ smm.msa oam.am soma om om~.mq mmm www.ma omm oom.mm mmm.mm mam.ma moma ammao spam mma.mm mmm omm.mma mam osm.~m oqo.om oom.mm coma :aaumo maq.mma.m mom.oa aoa.s~m.a mmo.ma mmm.amo.~ Nqa.msq.a oma.omm aoma .m .mmmo omm.aao mam.m smm.mmm ass.m mea.aas mms.amm «am.mma mama .z .mmmo mma.mq amm aom.ama New mmm.mm ama.mm mmm.aq moma momam umoo zoom mMo< zoom saw .xom .umm .xum amooa .Mom mama coaumucmesoou ucmEuwwuu puma mumoo omumaoomm< wumoo amusuosaum Ucm mama uomhoum .mamaaoa ca mama .moummma .muomaoam 6mm am :mwacoaz Na .mumm “moo nomaopm .a oaomH .xaccmaa< 307 .m nouqmno ca vocamaaxo mm .mumu Amom no: .mumuom coaum>pomzou cum coaumnaaanmum amusuasoaum< Qua: uuooom ca oaoe ma mmNa anocooo Mu< umcouuonm one .muoomoua ucooou you uaooxo .Mom How umcu ou acado omwa anocooo cm woesmmm mom Amumoo Hmuaamo woumwoommmv Mu< pom .sumcoa coauwa coaumsam>o uoomoua onu ma was» .owaa uoonopa map mocaeuwuoo mafia anocoom Mom mane .omwzponuo wamom om can .aouucoo cooaw pom zaaamEapa mum ucoeo>opnfia mo mxuoz onu MN .mumoh ooa mo uGoEuwo>cw ocu How wwaa anocooo cm woesmwm mom Amuwoo amuaomo ampsuosaum uoonoumv Mom pom "mo>aa owEocooo ucoEumo>ca amuaawom .wcman xHo3 mom cam moaam mom Ga .wucoESUOU uoomoua .mum .o Am.Nv moumu umwamuca coaumucmasoou couch umwuoucH nmumu mama pcoEo>ma£om mm cam Amumwkv mUONHoQ coaumuaupoe< cum cams pownopm o.n.m.muco oumm unoEo>oano< mm wow .mCONuquQEOQ Ammv muamocom unoEoocmncm pom vow muouomm coaummaupoe< .ucoo amuaamo pom mom as com: moumm pompoucH .mqu oqu anocoom ucoeumo>cH amuaamu .N macaw .XNucood< 308 .om-aa u u whom» uw>o Nos chowuwovm am now: .OHIH n u whom» pm>o Now "cowwHSumw .xao>wuooamou oa-ma anabaoo mom Nma .Nom .mosa coco: 89mm pom coco oau co comma mam couch Hmsccm u.m.am «comma mam mouma am scans coo: .musa mafia-oasumma now mum cacao mumc ouma ucoao>oanom “xoono HHNZM Non coo .Noa.. amnccm u N .OHIH n u whom» po>o cocoa ma NON umcuocm “Noe ma 0 n u now wumv mumu ucoEo>mw£om ”chm .uom: mm: q GESHoo ca czonm ouch may umcu wumuw madam xuoz omnmuoums mum .uo>ozom .oumu mfinu co comma mam: mcofiumusano mom .ma canaoo Ga ozonm mm ma ma ouch .:0aumucoE:ooo mom :0 owmmmo .m Houawno aw ponunam vmcwmaaxm ma mane .opou oEHu um coupsooo mono MN mm counsoo mum cowuma acaumaamumcw noonoua onu wcausu wcwnasooo mzoam ammo Ham umnu wcwcmoe .cowumaamumcw-ucwumcw oESmmm moumu moose .muomHOHa umoe pom nowwau manna cam Nalm mCESHoo cw couch onu umnu ooauoz .ANON + NmN + Nom n Nmmv Nmm um moamep much can om-ma whom» you .Aumoh you NN .ma-o mumom ao>o mucoEwuocfl amsccm amsvo cuwzv ma amok an NON Hmcowuwucm cm can .Aamoh poo Nm .mla munch no>o mucoEmuocH amsccm acado nuazv m pom» an NmN nonuocm .ouon mafia um Nom "coauma coaum5am>w paw» cm can wcapsc mzoaaom mm osuoom pasoz oEoocN Show mm mo cowuuoa mDAHz wAu umnu ooEdmmm mum uommouq wwnu pom .cowumaumSHHN Rom poomona xoouu xomHm can now mumu magma cocamaaxo on hoe wanna cam Na-m mcadaoo cw axonm camp on& .AmmmmHo cam amazhv muamocon poomouq mo mus“: uoauo ou no: cam .mm ou haco magma momma ucoEo>oa£om amsccm moose .ucmumcoo Cacao“ mono cows: noumm .Uownoa unmEQ0ao>oo mm onu wcflpso wmmouocw Au.mN.amv mouma ucwEm>oacom amsccm onH .mumo% amsoa>woCH How woumu unwEo>oanom oEoocNIEumm mm ouSQEoo on m Houamno cw cow: mum Ana-m mcEsHoo aw czocm mmv dump moose .moasooooua mom onHom co m uwuamno ca comamaaxo ma comp omocu mo om: "moon oump uno8o>oanom Ammv muwmmcoo acmeocmscmo .mOaumu umoo uwwocwn cousano-mum ocu moumEonuaam umnu usauzo nouDQEoo oosuopo ou Uoccopcw mum umnu mcowumHSEpom amoaumEonumE ow m nouamno aw com: mam hone .mmpsuoooaa mum mCNSOHHom m umuamno CH conflmaaxo ma couch mmocu mo om: onH .Ammv muamocmo ucoEoocmnco ouSQEoo ou mom an com: mm: Ana .aoov my oumm .ANMQfluwmcom Ca kamsoocoauo com: mm3 oaa.omam .mmmzm .hmswwumwzn .ocm.NNm .muamocon coaucouoou noom.Nm .muamocon uGoanao>ooon mo 85m moco.mNm u mmmmeo .coowasumw .mamm .coooz-oazm «mmao .am-omzm momm.ao .saaz-omzm «NmN.ao .amuou . .mom.oaam .aooou-ooa ammo.oum .coooz-opa «www.mmc .maoa-masumom-opa “02a .mcccoo assoc“ um: .oszc 093m .AmmmmHOV mmmm acscm muamocmo ponuo vowmaooomc: .mpomoaonu “www.mw .muamocoo quEmocmncm mon .mmnzm «cam.oaw .owumu o\m cou:a&ooumom ca poms mm .amuou .muamocon Havana .oauquo .umom non oqaw mo mmoa cannon um: m vomsmo mmpom co>oEouuuownoaa HN .GHHumo .MQN.oam .3oam poumzxomn mo Honucoo oooam on can .muamocoo owchmpo coawmaEN mommao ..o .omooo U .awo.mm .muamocon unoEQ0aw>wooa u mmmmHo ..z .mmmon .oomo - maoo u oozm - omega u comm mo omega .xooamo .mcmaa x903 spa: mocwumflwcoo you coco nonumu .mowoapsa wCNEEmuwoua HouDQEoo pom cowauowoumo mum: coco oEom pom .mcmaa xpoz :N czonw quu no“: wamafioo uoc has comp «cowumucoEdoou swag xpo3 .mom .m ou m.aocu:< mocow¢ w ommoao:a canyon um: m m cowumuco5500u oawN mafia mo owumm cauma o\m ammo-mom oazb pawocwp uCoEoocmccm mmmmao mmmzm cam mama uoomonm .co-mcma .coomaoam coo am cocaooaz Na .moom omomamm coo moacocmo .m canoe .xwvcona< 310 .oama .oa maoscmo .cocaooaz ea umaaocooo xuamm waaccmam mom .mmxo snow spas 3oa>uou2a cam mcmaa x903 oonmaoumz .mum .am maa>mmv uocmnoumz :uoncmm woaucsoo uoaumau cam coucaao .oonmuoumz Ho>am mamas-noon: umoz moaucsoo oommmzmanm cam uoauwau .coucaao .conmpmumz Ho>am magma-Homo: ummm Omma maaco .owmum 3oa>oa :a woman .moma Ga oumu coaumucoEDoou .muoomopm uo>am mammz mmma mm: .cmwanoaz .mucsou commonono .Uonmpoumz po>am xommaoauuaa moma HooEw>oz .cmwanoaz .mucsoo coumwca>aa .Uonmnoumz madam maaa>pmazom ”ca-Noma .coomaoum coo am cocaooaz “coco moma comm: .cmmanoaz .mucsoo ham .Uonmawumz comxoaHm-onoH moma once coma mamsacom .cmwacoaz .mucsou cowCaxoao .anmpmumz Hm>am commusum mo nocmnm ummm coma Cowmasum Noma mumsuomm .cocaooaz .mwaucsoo omaacmm cum pamao usamm .uoomma .Uonmumumz xooao aaaz mo nocmum cupoz Noma xoopo aaaz mmma aooouoo .cmwanoaz .mucnoo couCaao .conmuoumz xoouu umuxmsz mmma umuxmsz mmma mosh ..noaz .mmaucsoo oomocou cam oommmzmanm .3mcawmm .ownmaoumz xwouo Amsmoumaz mmma mmsmoumaz Ncma aaum< .cocaooaz .moczoo mmcaEOcmz .omompmooz um>am casuaa Noma causaa coma umooooo .cocaooaz .moosoo oooaoaz .omnmamocz cacao om ooma om mama maSh .cmmanoaz .mucsou dazvmao .cwcmuoumz camua coauxmopo Eumm moma xoopo Each coma mpmnanwm .cmwacoaz .Aucsoo coucaau .Uonmnwumz GOawmom-woaocmom muoumzncaaumo coma :aaumo acma .nom ..noaz .moaucsou awomoa 6cm omaacmm .vonmuoumz Hw>am ammo mo nocmam cusom acma .m .mmmu moma HooEoomo .cmmacoaz .mucsoo omaacmm .Uonwuoumz po>am mmmu mo nocmum wavoaz moma .z .mmmo mcma wash .cmwanoaz .zucsoo comm: .conmuoumz commz-xwouo xoMam moma xomam “cc-mcma .muocaopm omaosom Na oumo swam x903 nonmaoumz coaumucoESoov mom AuwaamV amCawaao on» now oumu coaumucoESoou Ucm .COaumooa cum cams uoomoum aasm pom mam: uoomopm .ONINmma .wuoohopm com am cowacoaz nocuo cam owaosum Na .mCOaumooa cam onmz .q canoe .xaucoma< 311 , . [IV-III ill-all. i. ll! : .mosa-mmm .Noc coo maaaz-mmm .Noo .mzzo .NOm cam omm .Nom .m ouoc mom .mamuou uomnoum ca noucsoo uo: wuamocon mumvcoowm . (Ill - II. ”wuawmcon usaOm powwooaam Honuzm onHo n "muamocmn una0n umumooaaw honuzm onhw .Amoma moan ou «mma ca Emuwoum mom am map mo :Oaumwoca map EopmV m .3 .moma ponouoo :.mcmam xaoz oocmuoumz com 3oa oaaosm cacao co maouom>oaz .mmmz .mmm .omm: .muoo .m.a “comam ago: .mom .omm: .oooo .noaz "mousom moo.mam.mmo u No.ooa No.m Nm.oa N~.a aNo.aa xNo.mo Nm.m No.m NN.~o .m.s amoo.moo.am n No.ooa aNm.NN Nc.a No.o Nm.cm aNc.mm No.o No.m No.~N .noaz ooc.maa.a mmo.o~m ocm.NN omm.o ooo.oao amoa.~ao mca.co Noo.oma oom.cmm .ouaz Noo.a cao.o mac.a -- -- oo~.m mmazo omNuo -- on caooomo maa.c moo.a -- -- -- -- moo.N oNN.a -- woma.o xocao .a moo.m oao.o Noa.a -- mom ema.a omma.m -- -- cmmo.~ Hmaze-a omm.o~ oom.oo ooc.ma -- -- omo.am omm.om -- omo.aN acm.m coma oaa.ma mam.mo oNa.N ooN.~N oom.~ ncoo.ma oo~.a -- -- comm.a .ccasoo ommnoc oNoucoN oomnom -- -- Noo.ooa oma.ooa mmo.mm NmN.Nc omm.oo aaaz coo m Now ca mam m -- -- cam.o mmc.0a omN.N mom.o -- .oaxcsz mmanoo coonomo ~o~.oma -- -- omo.~o~ oaa.coa a-- -- oaa.coa .cmomaz moN N cam ca -- -- -- omo.o omN.c oa. . -- o Boa ooc.oa omm.oa mam.~ -- -- oooa.o oooa.m om- m ommm m -- a om oom.~ ooN.m oom -- -- ooNc.a oonc.a -- -- -- apmm oaa.o ooo.m oom.~ -- -- ooo.o ooo.o -- mac moos.m :aaoco «m G. q III III a a a q u a mom. m mo .mma aam co . omm mm Noo.mm amo ca Nmm om coco co .m .oooo coo ma aom No Noc ha -- amo m coco mm oooo mm -- -- -- z coco mom.o Nma.ma -- -- -- oaa.N amo.c ocao.m mmao.m -- xooam cameo coaum Became acuoa coma oaaaz omega amsccm .w>< namuoe naumucooom louoom Iaw>ocom m23< Ammmv muawocoo coauco>oum vooam oEmz .mamaaon :a moon .Uowauowmumo .muamwcom uowmonm com am cowanoaz cam .m.b .m canoe .xaocomm< 312 .Nm.mm mo aouoo “Nm.a ca ceauooauaa-mzzc coo Na.om ca moccaopo-mzzo .ouoo cocanoaz so: .No.m ca zeauocaaua-mzzo coo .No.c ca moccaopo-ozzo .oooo .m.: on» go: .amuou NOOa can :a nou::oo on: wwwmu:oopoa am:oauaoom moonH .Nm.o .pwnuo c:m «Nm.o .:oaumopoop amu:ocao:a «mamas: Mono: amauumso:a u:m amaaoa:se .Nm.a «ousao:a uo: moo: .Nq.aa .owmu:woaoa mzz< "mum: .m.Da .mmm ponuo pom Nm.a wmonao:a .No.wo .owmu:oouoa mmm "moon .m.:x .mwo.mqq.am \ qu.mNmm caumu on» :o comma ma Nm.NN .mmm::oouom muamo:on mamo:oomm :mwanoaz one .mamuou powwow: oumammom on: :a woosao:a no: qu.mwa mo muamo:wn mumo:oomm mou3a0:an .o woo: :a cm:amamxo mm .Hocusm on: an dogwooaam momma on: o: w:e:aoo map :a :0: won .:E:aoo man: :a nou::oo .Nw.N no .mwm.oqm .muamw:mn u:a0n woumooaam movzao:aa .mzz< owm:amao o>mn wuoonopm :mwanoaz nonuo aam mmzz< :oaumwappa :omwasum: .mmmzm :mnu: haumoa no aam oumum :wmanoaz mc-mmma .mouo: moaom .:oauo:coua an owmouom woumo>ams w:aoa>ao an mummmooo: ma consano .Accma .omcmp «.D.: .:ouw:anmmzv coma «moaumauwum ampsuasoapm< .m oumum :mwacoaz mc-mmma «mu:oesooc oaam-:a .mow .oao w.ma o.ma m.qa o.Na :ou momma Hawsm o.NN o.Nm c.aN o.ma .5: m:moo canaoo mum o.qm o.mq o.mm 0.0m .sn umonz m.No o.oc c.am o.oo .sn cacao you .muoo N.0a m.ma c.aa 0.0a :ou ommaam :uou c.ao o.qw o.mc o.cm .sn :amaw pom .:aoo 33 fit 2.5 mowmuo>m owm:ampo owm:amnc owm:amuo oumuw :mwanoaz poo: .omaw noon .monm poo: ua:D mouo mo-moma -oooam -oooam mcaoooam occuo>< compo sccaooaz mo-mmma cam comaoam ammuo aaaz .ooom camam .o canoe .xaosmmmc .cclqcma How ma.Nm Us: mclmmma How NN.Nm .quwE%mQ uo< pmwfim mODaoxw mOOaHQ oncme US: Z< o .wom an com: moan: accamauo was scum owumsaoc . n .A.>a:: mumum .noaz ..coom .oa:w< mo .ummm .uzsamom mom “moasocv coaam :aoaw x N.o u moaam mmmaamc .Soma .88: tom :53; 982 of 5 523 .8,.“ 38:8 .55 reeoc .28m “505 .3398 ..m.: :a :.wannoum u:maum:no< o:m caco:a .oanm<: .xooaaz umuamz "Amaduwoam .u>ow mo am>oEmu cuazv coma pom w:0auoonoua w.3oo:mum .chma .cN coma: .mom u.o.a ..nmmzv m ooauoz ocano mano:oom .mom . HomV >2 wou:oanm::m «Accma aaam< .aao::oo on: “.o.n ..nmmzv . . . woamo:mum ooapm Eanou:a .aao::oo moousomom “mum: ..m.D "wooaum z: .mupwm w:a::wam :mManoaz .mom . Ham .nmmma .oau om.oa Na.aa oa.ma -.ma oo.oa coo cocoon aocso cm.N oc.m mw.m qq.m oc.m .sn m:mon wanauo mum no.0 cN.a «c.a mm.a mo.a .sn woos: ao.o mo.a oc.o mo.a om.o .ao cacao no: .muoo ca.o mc.m Nq.w No.5 nwa.oa :ou mommaam .:pou mm.o co.a ma.a mo.a oq.a .59 :ame you .:uou asoocoao z< co-ooma mc-moma yam can: mono .:wwa:oaz pom .mumaaon :a mooapm mono o>aum:noua< .m oaomH .xao:omm< 315 .ao>oa Amc-mmma uo:v mmma onu Eopm maoeamm an co>aooou mooapm mo xoc:a ow mo uoommo ozu o>oEop ouv mcma pom m:0auoomoua m.3oc:mpm nmsonuam .:zonw mooaua mclmmma on: mo :oauoscop NON w :o comma on: muonuo on: «zoo:mum Eopm hauowuao oo:amuno mum: wooaua :amhw acma: c:m .umo .:poo on: ha:Oa .m:0aumu:mEoo m.uonu:m “mooaum z< .m.O map umsncm ou mooan: mclmmma w:am: usn .o How mm osmmn .Ampmom macapm> .«OmO ”.0.0 ..cmmzv wooaum oanmuowo> o:w uazam smoum . .OmOmO “.0.0 ..nmmZV moauwaumum amusuasoapww m.aco mocom .noaz ..umoO mocoaom mono .oamom Eah coax m:oawm:omao c:m mumo mom :0 comma .oumEaumo m.ponusm .m.: qc-Ocma moan: cocao>o .noaz oc-ooma x .m.:z< n .ooazz< "ooosmeOo comm .mxmmv QHDUWNQ 300 “afloatbfl a.c.u:oov m canoe .xac:oma<